Nondestructive Characterization and Imaging of Wood
Nondestructive Characterization and Imaging of Wood
Nondestructive Characterization and Imaging of Wood
Editor: T. E. Timell
Nondestructive
Characterization and
Imaging of Wood
i Springer
Prof. VOICRITA BUCUR Series Editor:
Institut National de la Recherche T. E. TIMELL
Agronomique State University of New York
Centre de Recherches Forestieres de College of Environment Science and
Nancy Forestry
Laboratoire d'Etudes et Recherches Syracuse, NY 13210, USA
sur le Materiau Bois
54280 Champenoux
France
Cover: Transverse seetion of Pinus lambertiana wood. Courtesy of Dr. Cari de Zeeuw, SUNY eollege
of Environmental Scienee and Forestry, Syracuse, New York
ISSN 1431-8563
ISBN 978-3-642-07860-6
Bucur, Voichita.
Nondestructive characterization and imaging of wood / V. Bucur.
p. cm. - (Springer series in wood science; 760)
Includes bibliographical referenees and index.
ISBN 978-3-642-07860-6 ISBN 978-3-662-08986-6 (eBook)
DOI 10.1007/978-3-662-08986-6
This work is subject to copyright. AII rights reserved, whether the whole or part of the material is con-
cerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting,
reproduction on microfilm or in any other way, and storage in data banks. Duplication of this pub-
lication or parts thereof is permitted only under the provisions of the German Copyright Law of
September 9, 1965, in its current version, and permission for use must always be obtained from
Springer-Verlag Berlin Heidelberg GmbH .
Violations are liable for prosecution under the German Copyright Law.
© Springer-Verlag Berlin Heidelberg 2003
Originally published by Springer-Verlag Berlin Heidelberg New York in 2003
Softcover reprint of the hardcover lst edition 2003
The use of general descriptive names, registered names, trademarks, etc. in this publication does not
imply, even in the absence of a specific statement, that such names are exempt from the relevant pro-
tective laws and regulations and therefore free for general use.
Product liability: The publisher cannot guarantee the accuracy of any information about dosage and
application thereof contained in this book. In every individual case the user must check such
information by consulting the relevant literature.
VOICHITA BUCUR
Contents
1 Introduction ....................................... . 1
1.1 Brief Historical Review of Nondestructive Evaluation
of Wood and Aim of the Book ......................... . 1
1.2 General Concepts of Nondestructive Testing of Wood ...... . 2
1.3 Classification of Nondestructive Techniques for Wood
Quality Assessment .................................. . 4
1.4 Imaging of the Internal Structure of Wood ............... . 9
1.5 Summary and Outline of the Book ..................... . 11
United Kingdom by Hearmon (1965). Since the late 1970s, the activity of lab-
oratories all over the world on nondestructive testing of wood has been stim-
ulated by symposia organized in the USA at Washington State University by
Pellerin and coworkers in collaboration with the Forest Products Laboratory,
Madison (Ross and Pellerin 1991,1994) and since 1996 alternatively in Europe
and in the USA.
Since 1970, reference books have been published in the field of mechanical
characterization of wood (Jayne 1972; Bodig and Jayne 1982), dielectric prop-
erties of wood (Torgovnikov 1993), and acoustical properties of wood (Bucur
1995). Schniewind (1981) was the editor of the first encyclopedia of wood
science. The publication of an especially important series of books in the
Springer Series in Wood Science was initiated in 1983. The last 20 years of the
20th century have been characterized by an extraordinary development of dif-
ferent methodologies for nondestructive evaluation of wood products, which
has succeeded in the imaging of the internal structure of wood at both the
macroscopic and microscopic levels.
Due to the extensive literature on this subject, this book will provide an
overview of wood structure imaging techniques and the corresponding basic
concepts related to the nondestructive characterization of this material that
permitted the development of the modern imaging procedures. These out-
standing modern wood structure imaging techniques are a logical extension
of existing nondestructive methods developed previously and has benefited
from their breadth.
To promote the efficient use of wood materials in the future, three major areas
need to be addressed:
- development of nondestructive techniques for the evaluation of different
properties such as: physical, mechanical, chemical, aesthetic, etc.,
- improvement of natural qualities of wood through the modification of prop-
erties with different treatments, and
- creating new products using wood as a major raw material, corresponding
to the requirements of a modern society.
The development of nondestructive techniques has as it principal purpose
to reduce the uncertainty of wood products characteristics as influenced by
wood's biological nature. Despite the great attention given to quality control
in the development of manufacturing processes for glue laminated timber,
laminated veneer lumber or plywood, for particleboard and other wood-based
composites, interfaces are still the weakest link in the performance of these
products. To ensure interfacial integrity it is important to develop methods for
General Concepts of Nondestructive Testing of Wood 3
(")
0;-
t-1U!l1--t
'"'"S;
5 . T ()
A I O.3A ~
&W , .. 1 o·
::s
,,' o
...,
o ~ ~
::s
C>.
group of trees tree (m) annual ring (COl) cells (mm) Icell wa ll (J.lm)1 fibril (nm)l,cellulosic crystal ...'"~
'~"
~.
megascopic scale mesoscopic scalelmacroscopic scale microscopic scale submicroscopic scale
~
g..
::s
.Eo
Fig.I.I. Hierarchical structure of wood. (Data from Bucur 1995, with permission)
''"'""
...C'
~o
C>.
/::)
'~"
~
i!;"
'"
'"
8'"
a'"
V1
6 Introduction
Table 1.1. Nondestructive evaluation opportunities and the needs in the wood products. (Beall
1996, with permission)
Defects
Surface defects X X X X
Internal defects X X
Basic properties
Grain orientation X X
Density X X X
Moisture content X X X X
Strength/stiffness X X
Permeability/diffusion X X
Surface roughness X X
Geometry
Profile, shape X X X
Thickness X X
Bonding quality
Composites X X
Laminate materials X X X
Fasteners X X
Table 1.2. Classification of nondestructive methods for solid wood as a function of its physical
properties. (Sobue 1993; with perrmission)
1
(m)
tesy of Forest Products Society)
tltraViolet rays
JViSible light
]
infrared
microwaves
ultrasonic
waves
nuclear
magnetic +-_--'
resonance
Table 1.3. Scale of wood structural characteristic and the required resolution of imaging tech-
niques expressed by the wavelength
tecture of wood, the selection of the most relevant technique for structure
imaging is directly related to the resolution required by the selected method
and consequently to the wavelength, as can be seen from Table 1.3.
The general understanding of the interaction between the electromagnetic
waves and the wood material under inspection must be based on an accurate
description of the phenomena, with an increasing degree of sophistication of
the theoretical models. Consequently, the nondestructive evaluation of the
properties of the media becomes an inverse problem.
The development of a corresponding theory that describes the interaction
between the electromagnetic waves and the material requires three principal
steps:
- observation of the phenomena to provide guidance for the theoretical
approach;
- integration of the theory to the specific problem of observed phenomena;
- conduct of experiments that verify the first and the second steps.
The interaction between the second and the third step generates a model
that can be validated by the experimental data of the first step. The second half
of the 20th century has shown the development of a new branch of wood
physics related to the nondestructive evaluation of its properties.
in the future of low-cost devices will contribute to the widespread use of this
method. Neutron imaging is one of the most recent nondestructive techniques
developed for wood. High-resolution imaging (151lm) of water distribution
in wood can be obtained with small, clear specimens and with structural
elements.
The purpose of this introductory chapter, which is concerned with the nonde-
structive characterization and imaging of wood, is to provide a brief overview
of the material that will be discussed in the subsequent chapters. With com-
puted tomography two- and three-dimensional images of the internal struc-
ture of objects can be obtained without physical sectioning. Tomograms
obtained with different techniques, ranging from ionizing radiation (X-rays
and gamma rays) to ultrasonics and nuclear magnetic resonance and nuclear
imaging, provide excellent means of obtaining information about the internal
structure of wood.
The criterion selected for the sequence of the chapters is the wavelength of
the physical radiation, which is closely related to the scale of observation used
for wood characterization and imaging. This means that the book will begin
with X-ray and gamma ray imaging techniques, corresponding to the smallest
wavelength and consequently to the highest resolution, followed by thermal
imaging, microwave imaging, and ultrasonic and nuclear magnetic resonance
imaging methods. The last chapter is devoted to neutron imaging.
The book offers an in-depth review of the state of the art of the use of non-
destructive evaluation techniques as applied to the imaging of wood. The the-
oretical approach will be outlined, and emphasis will be placed on practical
applications. For the readers who would like to have more detailed informa-
tion about any particular technique, an extensive bibliography with a subject
index concludes the book.
2 Ionizing Radiation Computed Tomography
2.1 Introduction
of the specimen inspected with X-rays can be obtained (Polge 1966; Mothe
et al. 1998).
Because the high quantum energy of radiation and the low atomic numbers
of chemical constituents of wood, the attenuation phenomenon is caused
mainly by the Compton effect and consequently is proportional to the mass
density of the wood. It is generally accepted that the density of the cell wall in
oven dry wood is constant (I500kg/m3). Therefore, the density variations
observed on tomographic images are due to the distribution of various
anatomic structural elements of the specimen under inspection and to the
water content in the cell walls and in lumina.
The tomograms are obtained by calculation, using a sophisticated computer
program (Kak and Slaney 1988; Habermehl and Ridder 1994; Grundberg et al.
1996; Schmolt et a1.1999) that involves a complex technology. Each zone of the
specimen having the same density is evaluated statistically and is represented
by one color, or by a gray scale. More commonly, the red and the violet indi-
cated low wood density and yellow high density.
This chapter discusses the ability of X-ray or gamma rays to provide a non-
destructive method to identify and evaluate the internal structure of wood and
wood-based composites. X-ray or gamma ray imaging is a rapid tool able to
provide visual and analyzable internal structure of solid wood and of wood-
based composites. The ability of this method to quantify spatially the density
variations in solid wood, the presence of internal defects and inhomogeneities
in lumber or in wood-based composites, the variation of moisture content in
wood specimens, the distribution of water inside a trunk, etc. explains its
important increasing implementation in wood science and technology.
During the last 20 years, Scandinavian countries and Germany in Europe,
Canada, Japan, Australia, New Zealand and the United States have put stress
on the understanding of the basic aspects of X-ray computed tomography in
view of the development of specific wood, industrial scanners for routine
quality control of wood products (Benson-Cooper et al. 1982; Habermehl
1982a,b; Mc Millin 1982; Cown and Clement 1983, Asplund and Johanson 1984;
Qnoe et al.1984; Funt and Bryant 1985; Lindgren 1985; Harley and Morris 1988;
Davis et al. 1989; Wagner et al. 1989a,b; Rinn 1991; Habermehl and Ridder
1992a,b, 1993,1995, 1998; Grundberg et al. 1995; Schmoldt 1996; Sugimori and
Lam 1999; Qja et al. 2000).
The advantages of ionizing radiation computed tomography are numerous
when compared with conventional radiography. Firstly, this technique elimi-
nates the intermediate steps involving photographic film and optical densito-
metry and is able to make data available in real time. In process control and
in manufacturing situations, the density feedback is very important for tech-
nological competitiveness. Secondly, an important advantage of direct scan-
ning technique is the improvement of the calibration procedure, because the
mass attenuation coefficient can be determined directly, using a scintillation
detector with standard radiation pulse shaping and counting equipment, which
allows the user to select the energy range counted (Fig. 2.1). Thirdly, a large
Basic Phenomena 15
Fig.2.l. X-ray device for computed tomography on a tree. (Habermehl, pefs. comm., with
permission)
--"
rotation
"" "
~ translation
a)
slice
b) obj ect
y SD = X-ray path
r =ray of the projection
8 =angle of proj ection
c)
Fig. 2.2. Basic aspects involved in X-ray reconstitutive tomography. a Rotation of source-detec-
tor system around the specimen (Lindgren 1991b, with permission). b Volumetric element (voxel)
in a slice (Lindgren 1991b, with permission). c Beam path geometry between the source and the
detector. (Gilboy and Forster 1982, courtesy of Academic Press)
Basic Phenomena 17
As described in reference books (Herman 1980; Kak and Slaney 1988) and arti-
cles (Polge 1966; Hagglund and Lindgreen 1985; Lindgren 1985, 1991a,b; Fio-
ravanti and Ricci 1991; Habermehl and Ridder 1992 a,b; Lindgren et al. 1992),
in the case of monoenergetic photons such as y-rays or highly filtered X-rays,
with rectilinear propagation in solids, the relationship between incident and
transmitted ray intensities, the attenuation coefficient and the thickness of the
sample, is:
1= 10 e-I.lI (2.1)
where
I = intensity of the transmitted ray beam, passing through the sample
10 = intensity of the incident ray beam, passing through air only
/l = attenuation coefficient
t = thickness of sample
As ionizing radiation traverses the matter it is absorbed exponentially in
accordance with the equation:
1/1 0 = e 41t (2.2)
p = /ll/l' (2.3')
The density (Eq. 2.3') of the inspected sample is the ratio between the linear
attenuation coefficient /l and, /l', the mass attenuation coefficient, and can be
determined experimentally by measuring the difference between the intensity
of the incident beam and of the absorbed beam.
The Eq. (2.1) can be rewritten as:
Table2.1. Wood mass attenuation coefficient /1' (cm'/g) as found in the literature
The differences were calculated considering the data reported by Moschler and Marais as
reference
Basic Phenomena 19
The attenuation values calculated for each pixel are translated into a gray scale
or a color scale and transformed into a picture of the cross section of the
sample on a TV screen. For medical scanners, the CT number ranges between
-1000 for air to 0 for water and to +1000 for human bone. When working in
the optimum range, the medical scanners exhibit a typical photon limited noise
level of 0.5% in pixels of 1 mm2 with l-cm slice thickness. The image is often
reconstructed with 512 x 512 pixels and saved on floppy discs. In this way,
it is possible to use different digital image techniques such as geometrical
transformations followed by subtraction of the image to measure the dynamic
processes such as wood drying as a function of time or wood impregnation
with preservative substances as a function of time or of thickness of the
sample.
The relationship between wood density and the attenuation coefficient was
established by Lindgren (1985) for pine (Pinus sylvestris) on a Philips 210 CT
medical scanner and is expressed by Eq. (2.11) as:
Jl = 0.93 Po -1001 (2.11)
where Po is the oven dry density (kg/m3) and the attenuation Jl is expressed in
[H]. The corresponding correlation coefficient is r = 0.99 (Fig. 2.3). The rela-
tionship between the moisture content and the attenuation coefficient for
Po = 430kg/m3 and for Po = 540kg/m3 was also established as can be seen from
Fig. 2.4, using rectangular samples of 14 x 59 x 116mm and 30 x 97 x 250mm.
The accuracy was 1.5% for density measurements and 2% for moisture content.
It is well known that the attenuation measurements depend on different factors
such as: the energy level of the source, the moisture content, the geometry of
the specimen and the general noise observed on the tomograms.
H
r = 0.99
-520
-560
-600
-640
360 400 460 480 520
density kg/m3
Fig. 2.3. Relationship between oven-dry density and attenuation coefficients in Pinus sylvestris
sapwood (Lindgren 1985, with permission)
Basic Phenomena 21
Legend
j= 540 kglm3
H
CD H=3.ISa-506 r=0.99
@ H=4.71a-554 r=0.99
j = 430 kglm3
-100 <1l H=2.97a-609 r=0.99
® H=3.96a-646 r=0.99
§
~
::I -200
ii
~
-300
-400
-500
Fig. 2.4. Relationship between moisture content and attenuation for oven-dry wood at two con-
stant densities of 430kg/m3 in Pinus sylvestis. (Lindgren 1985, with permission)
Based on the data existing in the literature, Gierlik and Dzbenski (1996)
studied the influence of moisture content on the mass attenuation coefficient
for four levels of moisture content of wood, namely 0, 9.5, 18.5 and 26.8%. A
nonsignificant decreasing relationship was obtained.
The theoretical problems of the noise on the image were discussed by Gilboy
(1984) for the conventional translate rotate scanning geometry, using the fol-
lowing characteristics: D, diameter of the sample; N, pixel size; w, the width of
the pixel; S, ray sum; t, spacing between rays, usually the same as w; M, total
number of projections; n, finite photon count per ray sum; k, a constant
depending of the type of the filter used; T, total run time; /1, linear attenuation
coefficient; A', the effective activity.
The noise propagated to the image is estimated as a variation of the atten-
uation coefficient /1, and can be expressed by Eq. (2.12):
(2.12)
The first part of this equation, 4I1 kW. {D/tTA')1I2, depends on the geomet-
rical factors, the total run time and the source intensity. The second term of
Eq. (2.13), [sinh{J.1D)/{J.1D)3]1I2, is called "sensitivity factor:' and is accuracy
dependent. The plot of the sensitivity factor versus the diameter of the sample
to be inspected goes through a minimum and can indicate the choice of
optimum photon energy required for the best image.
Lindgren (1991a) identified the noise on five specimens of Pinus sylvestris
using a medical scanner. The noise in each pixel was ±4 CT numbers and a
variation with ±1 CT number corresponded to a ±2 kg/m3 variation in dry
wood density and ±6kg/m3 variation in wet wood density. No difference in
noise depending on annual ring width was observed as long as the same direc-
tion of rotation was used to obtain the images. The accuracy of the noise needs
to be established each time measurements are performed on the same sample,
for example, in studies of moisture content variation over time.
The image quality is dependent on spatial and density resolution and an
improvement in one can only be achieved by a reduction in the other. For a
given time of measurement, the use of a high intensity source (such as a gamma
source) reduces the compromise between these two parameters. Gamma
sources possess this important practical advantage and also have a constant
photon energy over space and time. X-ray tubes produce an energy spectrum
that is variable with time, due to supply fluctuations or to tube degradation.
They also induce variations of the take-off angle for fan beam measurements.
Beam hardening is affected by all these effects. Indeed, the effective energy of
the radiation is a function of many parameters related to a specific scanner,
and the corresponding resulting image is scanner dependent. Gilboy (l984)
noted "the lack of precision in the X-beam energy becomes a greater problem
when working in the photoelectric region where the strong energy dependence
magnifies the effect of small energy changes:'
Lindgren (l991a) noted that the total attenuation coefficient of a material J.1
is the sum of the attenuation coefficient du to the photo-electric effect, J.1P' and
of the attenuation coefficient due to the Compton scattering effect of photons,
J.1" as can be seen from Eq. (2.14).
(2.14)
where J.1 P and J.1c depend on the density of the material being tested, electron
density, effective atomic number and photon energy. Consequently, it is possi-
ble to calculate the attenuation coefficient for any material if the chemical com-
position and density are known.
The attenuation coefficients of principal wood constituents are: 0.2634 for
cellulose, 0.2655 for hemicellulose and 0.2608 for lignin. Lindgren (given cita-
tion) calculated the attenuation coefficient of wood and the corresponding CT
numbers for different volumes of cellulose, hemicelluloses and lignin and for
the oven-dry density Po = 500kg/m\ as shown in Table 2.2.
The decrease in the attenuation coefficient is related to the increasing lignin
volume and the decreasing cellulose volume. A more detailed analysis consid-
Basic Phenomena 23
Table 2.2. The influence of chemical constituents of wood on the attenuation coefficient and CT
numbers, calculated for an oven-dry density of 0.5 g/cm3. (Lindgren 1991a, with permission)
50 25 25 0.08775 -527
25 25 50 0.08753 -528
50 40 10 0.08799 -526
10 10 80 0.087l5 -530
Table 2.3. Theoretical mass attenuation coefficients of wood calculated for two levels of photon
energy and for the following mass fraction of elements: 50% C, 44% 0,5.7% H, 0.3% N (Tiitta
et al. 1996) (lOth International Symposium on Nondestructive Testing of Wood, 2002, Press Poly-
techniques et Universitaires Romandes, p. 189. Reproduced with permission of the editor. All
rights reserved)
ering the mass fraction of C, 0, Hand N was proposed by Tiitta et al. (1996).
They calculated the mass attenuation coefficient of wood for two levels of
photon energy and for moisture content ranging from 0 to 100%., as can be
seen from Table 2.3. The mass fraction of the elements was: 50% C,44% 0,5.7%
H, 0.3% N. From this theoretical approach, it was confirmed that the photon
energy level is a very important factor for the definition of mass attenuation
in wood. For example, at 0% moisture content, the mass attenuation coefficient
is 22.61 cm2/g at 5.9keV and 0.192cm2/g at 59.5keV. It was also noted that at
the same level of photon energy the mass attenuation coefficient increases
slightly with moisture content. For example, at 59.5keV and 10% moisture
content, the mass attenuation coefficient is 0.193cm2/g and at 100% moisture
content, it is 0.198 cm2/g. A possible explanation of this statement is the small
difference between the mass attenuation coefficient in water (0.205 cm2/g) and
in oven-dry wood (0.192cm 2/g).
The linear regression equations established between the attenuation coeffi-
cients expressed in CT numbers and the measured density of oven dry wood
and of green wood are given in Table 2.4. The proposed equations are calibra-
24 Ionizing Radiation Computed Tomography
Table 2.4. Regression relationships between wood density and CT numbers for different
moisture contents. (Lindgren 1991a).{with permission)
tion equations. They must be established for each scanner used for laboratory
or industrial measurements.
As noted in reference books and articles (Herman 1980; Kak and Slaney 1988;
Gilboy and Foster 1982; Gilboy 1984; Fioravanti and Ricci 1991; Schmoldt 1996;
Bahy11998; Coles 1999), image reconstruction methods are:
- analytical methods, based on Fourier transforms and related techniques
including filtered back-projection (convolution). Data obtained from con-
volution can be processed immediately as collected and the virtual image
built by projections is available instantly after scanning. Finite Fourier trans-
form (FFT) is periodical in the time domain, requires an infinite number of
projections and gives a picture in the frequency domain in the Cartesian
coordinate system. In practical situations, the number of projections is not
infinite as it should be. The Radon transformation in the image reconstruc-
tion process overcomes this aspect. Convolution of the density function
using a filter function realizes the filtering process of the transformed data,
for a selected angle and position. A huge amount of computation is required
with parallel projections or with fan beam projections. The effect of an
increasing number of projections on image quality has been reported and
seems to be limited to 45 projections, as noted by Gilboy (1984) for a spec-
imen in light alloy casting. When data rates are low, this method can give a
useful saving in measurement time.
- algebraic (iterative) methods need a very large amount of data for image
reconstruction, and better results are obtained from relatively few
projections.
Each attenuation value used for the image reconstruction is associated with
a specific cross-sectional pixel size and finite slice thickness (creating the
voxel). Pixel squared size is between 0.9 and 3 mm. Pixel resolution depends on
the source-detector distance and on the number of detectors. Slice thickness
can vary from 1.5 to 20mm (Schmoldt 1996) and limits the sensitivity and res-
Basic Phenomena 25
1000
<I>
0;
'"
'0.
'-
800
0
600
400
200
0
0 40 80 120 160
b) gray level
Fig.2.S. Histogram of the image of knots and cracks in a red oak log (Faust et al. 1996). a CT
image of red oak section; b histogram of the image and threshold selection. (10th International
Symposium on Nondestructive Testing of Wood 2002 Press, Poly techniques et Universitaires
Romandes, p. 203. Reproduced with permission of the editor. All rights reserved)
were incorporated into the model. This was a very original application of this
model.
Faust et al. (1996) presented a flowchart for a system (Fig. 2.6) to detect knots
and cracks in red oak, black walnut and hard maple, with 3D analysis of the
position, orientation and shape of these features. The basis of this analysis is
the density. The system first processes each single image and then relates
processed results of each CT image across neighboring CT images to pinpoint
defects. The most current algorithms link annual ring structure in an edge-
detected image. Locating the pith is an important step in identification of
Basic Phenomena 27
selective smoothing
yes defect-like
~--+I region list
Fig. 2.6. Flow chart for the detection of cracks, knots, earlywood, and latewood in a hardwood
(Faust et al. 1996) (lOth International Symposium on Nondestructive Testing of Wood 2002, Press
Polytechniques et Universitaires Romandes, p. 202.Reproduced with permission of the editor. All
rights reserved)
growth rings and the longitudinal axes of knots and cracks that normally pass
through the pith. Table 2.5 shows the percentage of images with pith location
precision with different numbers of pixels ranging from 5 to 20. The best loca-
tion was obtained with 20 pixels for walnut, a semi-ring porous species.
28 Ionizing Radiation Computed Tomography
In 1979, the works of Cormak in South Africa and Hounsfield in England were
recognized with the award of the Nobel Prize for Medicine for the develop-
ment of ionizing radiation computed tomography. Cormak (1963) manually
scanned collimated beams of 60Co gamma rays across aluminum and wood
phantoms and detected the transmitted photons with a GM counter. He recon-
structed the image of the scanned sample from one-dimensional projections
viewed at different angles in terms of a map of linear attenuation coefficients,
using a mathematical model he developed. Hounsfield (1972), using a much
more intense X-ray source, produced images of a human head in a run time of
several minutes, using a one-line minicomputer. This was the starting point for
the development of X-ray computed tomography for medical use. It is also
notable that today the unit used for X-ray imaging with medical scanners is
the Hounsfield [HJ. The Hounsfield scale for medical scanners is between -
1000, corresponding to attenuation in air, and +1000, corresponding to atten-
uation in human bone, and is 0 for attenuation in water. This scale is linear
between 1000 and 0, and its accuracy is ±2 H units.
Medical scanners were used in the 1980s for nondestructive inspection of
different nonmedical objects, and new devices were developed for specific pur-
poses. Medical scanners were first used for imaging wood specimens because
the physical properties of wood do not differ greatly from that of the human
body. The energy normally available on a medical scanner is about 70 ke V, with
soft X-rays. Presently, in wood science and technology, four generations of
scanners are being used. A detailed description of each type of scanner will be
given in the next chapter.
The development of the nonmedical scanners, first experimental and later
commercial, started in the mid-1970s (Reimers et al. 1984). A wide range of
applications were emphasized, ranging from an investigation of the foaming
behavior of various coals at high temperature, concrete specimens, cracks and
delaminations in metallic structures, defects in ceramics, fluid distributions
inside porous simulated rock cores, wood structures and trees, thick plastics
to a maximum 400mm, crack detection in welded steel bridges, internal fea-
tures in jet and automobile engines, fluid flow measurements in nuclear engi-
neering, to a complete inspection of the Trident C-4 rocket motor which is
nearly 2 m in diameter. However, the cost of the systems is a real barrier, and
the choice of testing methods must be made chiefly on economic grounds. The
industrial systems in the wood industry require high energy X-ray sources,
accommodation for a wide range of specimen sizes and densities and opera-
tion in industrial environments with a speed imposed by effective wood
processing.
The major system components of an X-ray scanner are: X-ray generation
system, fan beam or parallel beam X-ray collimator, system for data collection
and analysis, image display software, mechanical gantry with possibility of
Equipment for Imaging Techniques 29
Fixed or portable equipment has been used for different purposes in wood
science and technology. If the test object can be brought to the scanner, which
is the case of logs and boards, fixed equipment is recommended. In this case,
an X-ray source has commonly been used. The portable equipment was
designed to be used for in situ inspection of trees, poles and building elements.
Gamma rays, which are mono energetic, were used in this case because they
have many advantages in the wide range of energies available, portability and
cost. In spite of these superior characteristics, gamma ray sources are of
low intensity, which is 104 times less than X-ray tube sources (Gilboy 1984).
Malan and Maris (1992) have summarized the use of gamma rays for wood
densitometry.
a)
"
detectors
X-ray source
c)
Fig. 2.7. Scanner system with detectors perpendicular to the axis of rotation (Schmoldt 1996).
a First generation, a single detector, translate-rotate system (fan system). b Second generation,
multidetectors, translate-rotate system. c Third generation, rotate only.
Fig. 2.7. d Fourth generation, stationary detector, rotate only (with permission)
located on a circle that surrounds the X-ray source and the sample. The main
advantage is the simultaneous data collection through the entire sample for
each view, resulting in fast scans. Due to the very high cost, the system is rarely
used.
To avoid the limitations of all four generations of scanners, an alternative
was proposed with a scanner operating tangentially as can be seen from Fig.
2.8. The main advantage of this system is its ability to collect data for the entire
volume of the object simultaneously from many cross-sectional slices. Most
images are generated using data from a single detector. The speed is faster than
that of the third- and fourth-generation systems. The mechanical motion
system is very simple. The limitation of this system for the moment is the lack
of a fast reconstruction algorithm.
The capability of the tangential system to generate images of any planar or
curved surface within the object was demonstrated in laboratory work by
Lindegaard -Andersen et al. (1990) for a block of pine wood of 60 x 60 x 51 mm.
The distribution and shape of the annual rings were clearly identified. The
detector array is placed parallel to the axis of rotation of the sample and per-
pendicular to the cross section. A fan-shaped X-ray beam is formed by the
source and the detector array. The number and spacing of detectors determine
how many tomograms can be collected simultaneously.
32 Ionizing Radiation Computed Tomography
"' .
Fig. 2.S. Tangential scanning system. (Schmoldt et al. 1999, with permission)
Qualitative and quantitative information can be obtained from the X-ray com-
puted tomography images. Qualitative applications require only visual assess-
ment of the images and include evaluations of internal flaws and orientation
of the samples. Quantitative applications of X-ray imaging are obtained in
Equipment for Imaging Techniques 33
Table 2.6. Optimal X-ray energy, combined effect and transmission probability for softwood
species (1 mm thick) (Olson et al. 1988, courtesy of Wood Fiber Science)
addition to qualitative ones and are related mainly to the measurement of the
density, moisture distribution and porosity of the wood. To obtain accurate and
reliable data from X-ray imaging, the scanning must be carried out in a very
consistent manner as was demonstrated by Lindgren (1985, 1991a,b), Liu et al.
(1988), Moschler and Dougal (1988), Olson et al. (1988), Fioravanti and Ricci
(1991) and others.
X-rays travel in a straight line through materials. The monoenergetic beam will
radiate perpendicularly on a piece of thin wood with uniform thickness and
heterogeneous density structure, consisting of four major elements (R, C, N,
0) and various ash ions (Ca, K, Mg, P, Si) present in minor quantities
(0.15-0.37% based on oven dry weight). Two main parameters (Table 2.6) are
related to the beam path as defined by Olson et al. (1988): the range of trans-
mission probability and the combined effect. The range of transmission prob-
ability (AP) is defined as "the radiation resolution of a given wood X-ray
experiment, in reference to the separability of transmission probabilities
resulting from density variations in the radial direction of wood:' The com-
bined effect C is a global factor taking into consideration the combination of
X-ray energy, wood specimen thickness and heterogeneities induced by the
anatomic structure. Consequently, the maximum radiation resolution is deter-
mined by density variations in wood specimens, X-ray energy (keV), X-ray
intensity (rnA), spatial and contrast resolution, display contrast, and beam
hardening. The quality of the tomographic image is directly related to the X-
ray energy. Because X-ray attenuation is highly dependent on the energy of the
X-ray used, it is always necessary to provide this information for laboratory or
field measurements. Theoretical studies published by Olson et al. (1988) have
34 Ionizing Radiation Computed Tomography
1.00
g 0.80 LEGEND
:g
.n
0
....
p. pine heartwood
=
0
'in
0.60 t= 1.0 mm
·s H = heighest
en
en L = lowest
§ 0.40
p
0.20
0.00
I 2 5 10 20 50 100
photon energy (keY)
Fig. 2.9. Transmission probability and X-ray energy in pine heartwood of 1 mm thickness (Olson
et al. 1988). H Highest value; L lowest value. (Courtesy of Wood Fiber Science)
spruce heartwood
10- 1
10- 1
photon energy (keV)
Fig. 2.10. The influence of photon energy on mono atomic attenuation coefficients for C, 0, H, N
and for spruce heartwood. (Liu et al.1988, courtesy of Wood Fiber Science, with permission)
Table 2.7. The influence of CaCO, on the attenuation coefficient of wood and on CT numbers.
(Lindgren 1991a, with permission)
Cellulose (%) Hemicellulose (%) Lignin (%) CaCO, (%) flwood CT number
45 30 25 o 0.0878 -527
45 29.5 25 0.5 0.0882 -525
45 29 25 0.0886 -523
Spatial resolution is the ability of the system to distinguish two or more closely
spaced high contrast details of the internal structure of the specimen under
test. The spatial resolution of any imaging system is controlled by the accuracy
of ray paths through the sample and is related to the following factors: opera-
tive size of the source and detector, distance between the source/detector
and the sample, number of independent projections produced by the system,
ability to separate overlapping features information, signal-to-noise ratio
at small source and detector apertures (usually <0.25 )lm), and the cost of
computation.
36 Ionizing Radiation Computed Tomography
Table 2.8. The size of the pixel for the array of 512 x 512 as a function of the size of the FOY.
(Fioravanti and Ricci 1991, with permission Holzforschung)
(pix = FOV/512)
Equipment for Imaging Techniques 37
ric resolution. For wood, the intensity can be higher than (130 rnA) for a scan-
ning time of 3 s. Fioravanti and Ricci (1991) published an example of the matrix
of density (Fig. 2.11) with all the measured values for all voxels, given in a
matrix of 11 lines x 5 columns. A specific voxel can be identified with the recon-
struction matrix of 512 x 512 elements. The visualization with points is a
concise representation of all experimental data, making possible the observa-
tion of the isodensity curves inside a selected annual ring (Fig. 2.12). The pre-
cision of the measurement of the attenuation coefficient, using X-ray computed
tomography, is ±0.5%, compared with conventional radiographic techniques
for which the precision is ±3%.
TabIe2.9. Relationship between the voxel volume and the slice thickness (Fioravanti and Ricci
1991) (with permission)
Fig. 2.11. The attenuation values superimposed on the tomographic image of annual rings
pattern. (Fioravanti and Ricci 1991, with permission)
38 Ionizing Radiation Computed Tomography
Fig.2.12. Isodensity zones in an annual ring. (Fioravanti and Ricci 1991, with permission)
Theoretically, density is a scalar and not a vector and a priori is not influenced
by the anisotropic direction of wood. Nevertheless, because wood is a natural
composite, laboratory studies have reported the influence of the anisotropic
direction on the measurement of attenuation coefficient (Moschler and
Winistorfer 1990; Fioravanti and Ricci 1991; Malan and Marais 1992;
Karsulovic et al. 1999). For precise densitometric measurements, it is
recommended to orient the irradiation direction of the specimen with the
longitudinal anisotropic direction of the wood. In this case, the number of
photons absorbed by the cell wall is minimal. If scanning along radial or tan-
gential directions is required, it is obvious that the scanning inspection must
be performed parallel to the selected anisotropic direction. The attenuation
coefficient depends on the thickness of the cell wall. Errors in attenuation mea-
surements are created if the specimen is presented in an inclined position
versus the incident radiation, because the experimental wall thickness is
greater than the theoretical.
Table 2.10 gives the effect of anisotropic direction on the linear attenuation
coefficient measured on 141 wood specimens {Malan and Marais 1992}. The
longitudinal values compared with those in radial and tangential directions are
the lowest values. No significant differences were observed between the radial
and the tangential coefficients. The linear relationships between the density
and the linear attenuation coefficients in radial and tangential directions for
Pinus radiata are given in Fig. 2.13. The mean value of the attenuation coeffi-
cient in the radial direction was 4% greater than in the tangential direction
but the difference is not significant at the 95% confidence level. Moschler and
Equipment for Imaging Techniques 39
Table2.10. The effect of anisotropic direction on the linear attenuation coefficient. (Malan and
Marais 1992, with permission Holzforschung)
0.14...-----------------------,
0.13 radial
irradiation
S 0.12 tangential
-2 irradiation
::, 0.11
...l
:::1. 0.10
0.09 .j. _ _ _
knot
~-/_-..r
0.08 H-HH-+-t--H,-t-t-t-H,-t-t-++t-t-t-++-t-HH-+-t-H,-t-t-+-t-t-t-t-+-H
o 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
point measured on Y axis (em)
Fig. 2.13. Relationship between the density of Pinus radiata wood and linear attenuation coeffi-
cients in radial and tangential anisotropic directions. (Karsukovic et a1. 1999, courtesy of Forest
Product J)
Table 2.11. CT numbers of different materials. (Hattori and Kanagawa 1985; courtesy of The
Japan Wood Research Society)
2.4 Applications
X-ray computed tomography opened a new era for the nondestructive exami-
nation and inspection of trees and poles. The quantitative determination of
decay, knots, checks, heartwood formation, ruptures, wet core, spreading of dry
decayed tissue and other major defects as well as the distribution of moisture
inside the stem of a tree is now possible. Fundamental differences among the
images obtained for different species are easy to recognize.
Pioneering research results related to the development of portable instru-
ments, suitable for outdoor use were published by several authors (Habermehl
1982a,b; Onoe et a1. 1984; Habermehl and Ridder 1992a,b,1993, 1994, 1996;
Niemz et a1. 1998a,b). A description of these devices will be presented in the
following pages.
For the examination of forest trees or of park and street trees, Habermehl
and Ridder (1998) proposed two types of portable instruments, one with a par-
allel beam system (Fig. 2.14) and another with a fan beam system (Fig. 2.15),
equipped with a y-ray source. The source and detectors are sustained by the
mechanical device shown in Fig. 2.16.
l:?
o
'0
E laptop computer
Fig.2.14. Portable apparatus for the examination of tree with a parallel beam system.
(Habermehl and Ridder 1998, with permission)
42 Ionizing Radiation Computed Tomography
measuring plane
within the tree
bearing ring
source of radiation
Fig. 2.15. Portable apparatus for the examination of trees with a fan beam system. (Habermehl
and Ridder 1998, with permission)
Fig. 2.16. Mechanical device for the sustaining of the source and detectors of the portable appa-
ratus for the examination of trees. (Habermehl and Ridder 1992b, with permission)
The size and location of the defects that can be observed on a tomographed
cross section are cavities and decay of the trunk, mechanical injuries, fissures,
holes, zones destroyed by fungi, areas of lowered density with less moisture
than in healthy zones, the presence of metallic inclusions in the trunk, like nails
Applications 43
and bullets, zones of perfectly healthy wood, and the difference between
sapwood and heartwood in some species because of the difference in their
moisture content. The devices are very effective for the examination of park
and street trees. The decision of the environmental authority to conserve trees
or not after examination of tomographic images can be objective, and the
safety of the traffic on the streets can be assured. For forest trees, the questions
answered by the portable apparatus are related to the effect of liming
and fertilization on the sapwood to heartwood ratio, to the water pathways
above the cross section of the trunk, to the existence of frost cracks, and to the
existence of wet cores and other specific internal structures that can affect
the variation of the attenuation coefficient of the radiation and consequently
of the density.
The source of radiation is 137 cesium having a half-life of 30 years and emit-
ting y-rays of a quantum energy of 662keV. The parallel beam system is
equipped with a source of radiation of 13 GBq and allows the translation and
rotation of detectors and only translation for the source. The detector system
is consists of three detectors with an angle of 3.75° between them. One trans-
lational displacement produces three simultaneous attenuation profiles at the
same time. A total of 48 profiles are recorded in 16 rotational positions. The
smallest diameter of the tree to be measured is 12 cm with a counting rate of
19,000 counts/s and the largest diameter of the tree is 72cm. The tomograms
reconstructed on the computer are performed in less than 2 min for one profile.
The reconstruction process is based on the analysis of a matrix that contains
values of the attenuation coefficient (cm-I ) of the examined slice. In the case
of a tree, the reference value was considered to be 83 units for water and 0 for
air. (Hounsfield units in medicine are -1000 for air and zero for water.) The
spatial contrast for the standard adjustment is 10 mm for the full width at half-
maximum (FWHM) of the signal. For trees with diameters up to 48 cm, the res-
olution must be 7.5 mm FWHM of the signal. The contrast resolution obtained
was 2% and was related to the water attenuation.
The fan beam system is a more complex system and has 30 detectors. The
source is 185GBq and 8000 counts/s, mounted on a metallic ring with a 100cm
diameter. The complete data set produces 150 profiles, measured with a rota-
tion of the source-detector system with a step of 2.4°. The fan system has the
same spatial and contrast resolution as the parallel system. The algorithm
reconstruction for the PC time is 4 min for one profile. For larger trees the total
examination time is about 1 h. The tomograms can be coded by color or by a
gray scale, as can be seen in Figs. 2.17 and 2.18.
Using computed tomography Onoe et al. (1984) were the pioneers in the
development of an automatic method for ring width measurement and ring
counting of several species, including red pine, black pine, Douglas fir,
44 Ionizing Radiation Computed Tomography
a)
24
30
--
36
42
48
--_.. _---
54
60
66
72
78
84
90
96 24
30
------
36
42
48
54
60
66
72 ----
78
84
90
%
Fig.2.17. Tomogram on a color scale of a transverse section of a healthy (a) and of a decayed (b)
Norway spruce tree. (Habermehl and Ridder 1996, with permission)
Fig.2.18. Tomograms on a gray scale (Habermehl and Ridder 1996). Legend: a) a lime tree with
internal decay and big holes b) an ash tree, showing a metallic inclusion (with a very high atten-
uation) (with permission)
Japanese cypress, birch, horse chestnut and maple. The details of the annual
ring structure can be seen using high spatial resolution in the image recon-
struction with computed tomography. A few tenths of millimeter resolution
can be obtained with well-collimated, narrow X-ray beams and dense sam-
pling. The device designed and used by Onoe et al.(l984) was portable, with
an X-ray tube of 40-120keV and three NaI scintillation counters. A fan beam
algorithm was used for reconstruction, and 1200 samples of 16-bit projection
data were taken at 2° intervals. The reconstruction images show very clearly
Applications 45
the annual rings in Douglas fir as well as the heart wood central zone, having
decay (Fig. 2.19).
Figure 2.20 gives the profile and the width of automatically measured rings.
Firstly, the pith was determined by contour tracing of an inner ring. Secondly,
several diameters were drawn through the pith. Thresholding after differenti-
ation yields the number and the width of the rings. Hauffe and Mahler (2000)
studied the possibility of evaluation of log quality using only the width of
annual rings measured with a scanner on spruce logs from southern Germany.
The technique was successful for 80% of the logs. Rinn (1991) proposed a
mobile tomograph with 12 sensors for scanning growth rings in trees and
poles.
Fig. 2.19. Reconstructed images of the cross section of Douglas-fir pole compared with a photo-
graphic image of the same section (Onoe et al. 1984). a The reconstruction image of the wrap-
ping pattern of the annual ring around a branch. b The reconstructed image of the heart decay
in the central zone. c Photographic image of a. d Photographic image of b. (Reprinted from
Nuclear Instruments and Methods in Physics Research, vol. 221, 1984,213-220,2001, with per-
mission from Elsevier Science)
46 Ionizing Radiation Computed Tomography
c
--I-HlIIIHHHt-Hfff-lI-fIH1HIIHHHI-lUllIlIIIHtItHHlftlH1HHnUfHHHHlI----
Fig. 2.20. Profile and width automatically measured annual rings (Onoe et aL 1984)_ (Reprinted
from Nuclear Instruments and Methods in Physics Research, voL 221, 1984,213-220,2001, with
permission from Elsevier Science)
Another aspect of the examination of growth rate of trees where the appli-
cation of X-ray computed tomography is very successful is related to the devel-
opment of microtomography for 3D anatomic structure examination, as a very
important step in the further development of wood microdensotometry in
general. Davis et al.(l991) reported very well-defined images of transverse
structures of wood of Pinus sylvestris and Eucaliptus delegatensis, as can be
seen in Fig. 2.21. Details of the structure of the annual growth rings are clearly
visible, and tracheids, rays and resin canals can be observed on the Pinus
sylvestris specimen. For Eucaliptus delegatensis, a collapse check in the high-
density zone is observed. The device for X-ray microtomography of wood was
attached to an X-ray diffractometer. The schematic representation of this
device for X-ray microtomography is shown in Fig. 2.22. The technical char-
acteristics of this X-ray microtomograph are as follows:
- X-ray energy: 8 keY;
- beam characteristics: two beam collimation arrangements, one cylindrical
with 20 Jlm and another rectangular of 5 Jlm width and 100 Jlm height;
- translation movement 255 stepwise, followed by 1.50 rotation repeated 120
times for 1800 rotation of the specimen;
- reconstruction algorithm using a summation-convolution back-projection
method;
- data collection time: 7 h for 20-Jlm scan step;
- computation time for image reconstruction: 5 min;
Applications 47
Fig.2.21. Reconstructed images of the anatomic structure of Pinus sylvestris (Davis et al. 1991).
a 5-l.Lm spatial resolution of the transversal section (0.9 x 0.6mm) of the annual ring with early-
wood and latewood zones. b 5-l.Lm spatial resolution of the transverse section (0.9 x O.5mm) in
which the highest density is observed in the middle lamella between the cells. The images were
obtained with the first generation scanner (8keV, X-rays from the Ka peak of a eu target tube).
(with permission)
- pixel gray is zero (black) for the air and 255 (white region) for highest
density in latewood;
- X-ray attenuation coefficient is defined with 5% accuracy;
- spatial resolution 111m;
- specimens section 2 x 2 mm of Pinus sylvestris and Eucaliptus delegatensis;
- annual ring width: 0.4 mm.
.,..
ex>
RS232C
r-,
------·1 Link r - - - - - - - , ~i
interface controller ~ Micro-Computer! ), 0'
8.
(MC68701) System Graphic Display t::.
::;I
11v------/ .
k I and (JQ
r-----~I . i I
{l Hard disk I Ii ,Image Processing ~
!!;.
o·
::;I
stepper III----- n
X-Ray generator o
motor drivers .g
RS232C Link ;;.
(!)
0-
S'
y~ Ni filter o
3
(JQ
H r ~
X-Ra;tube I ~
I! (Cu ~/S\~ nl ' JI ~
Target) r n ~ n: object: .. jdetector Na(TI)~';l Data Acquisition
. U U\~~~/ ~ uI 1~lsystem
C1 C2 ~ C 3 '~- - - - - - - - - '
\ translate - rotate stages
Fig. 2.22. X ray microtomographic device. (Davis et al. 1991, with permission)
Applications 49
The moisture content of living trees is related to the viability of trees and is
dependent on many factors such as climatic conditions, the moment of the day
when the test of water flow is made, the sanitary conditions, etc. The forest
decline observed today seems to be induced by the disturbance in water trans-
port in trees.
The portable computer tomograph proposed by Habermehl and Ridder
(1992a,b) and used by Szendrodi et al. (1994) permits mapping the distribu-
tion of water inside the trunk and observation of the mechanism of water
transport in xylem. Raschi et al. (1995) and Tognetti et al. (1996) utilized a
portable apparatus equipped with a parallel beam system and with a source of
radiation of 13 GBq for measurement of the water content in the stems of
sessile oak (Quercus petraea) and in Turkey oak (Quercus ceris) (Fig. 2.23) . For
the portable computer used in this experiment, the CT numbers are on a scale
corresponding to 83 units for pure water and 0 for the density of pure air, on
a grid of 75 X 75 pixels for Quercus ceris. and of 63 x 63 pixels Quercus petraea
with a reconstruction interval of 3-4 mm. The tomograms for Quercus ceris
displayed green and blue colors, indicating a high density in the stem. The
sapwood had a high density (green color) and the heartwood had slightly lower
density (dark blue color). In Quercus petraea, the sapwood (blue and green
rings) is well hydrated. The embolized zone is represented by the violet ring.
The heartwood of highest density is represented by a dark blue color. These
species are characterized by different susceptibilities to drought; the first one
is a mesophytic type species while the second is xerophytic. The trees origi-
nated from a forest in Hungary at latitude 47°N, longitude 20° and altitude
320-340 m, with an annual precipitation mean of 600 mm. The trees from this
forest had suffered a prolonged period of drought for many years before the
experiments, which were performed in the spring. Characteristic anatomical
structures of these species are schematically given in Fig. 2.24, where large
vessels can be seen in the earlywood. The main function of large vessels is to
increase the efficiency of sap flow. In this tissue, cavitation risk is always
present. When the large vessels, malfunction, the sap flow is supplied by the
50 Ionizing Radiation Computed Tomography
a) b)
2". 00
,. ...
,..aD 2 •. 00
JO .OO
• . 00
-
11.00
Fig. 2.23. Tomograms of Quercus petraea and Quercus cerris (Tognetti et al. 1996). a Quercus
".00
cerris; b Quercus petraea. N, E, S, Ware the cardinal points; 1,2,3,4 are the position of measured
--
.....
N."
points The density scale in CT numbers is represented by different colors (i.e., CT = 30 corre-
sponds to dry condition and CT = 96 corresponds to high moisture content). (with permission)
o
o
I ,
Quercus ceris
(b)
o
Quercus pelraea
Fig. 2.24. Anatomic structure of a Quercus petraea and b Quercus ceris. (Jacquiot et al. 1973, with
permission)
Applications 51
smaller latewood vessels and fiber-tracheids, which remain functional for years
(Zimmermann 1983).
In both species, a narrow ring of low density surrounded the high-density
ring which indicates that the outermost vessels buffer the sap flow rate (Granier
et al. 1994). The sapwood area is influenced in different proportions by the sap
flux as can be seen from the different colors displayed on the tomogram. The
relative uniformity of the density in Quercus ceris compared with Quercus
petraea could partially explain the better drought resistance of Quercus ceris.
In Quercus petraea a high proportion of red and violet zones on the images
indicate a relatively low wood density. The active sapwood appears in pale
green and blue zones corresponding to 90 and to 72 units, respectively. The
tree examined was not very healthy because of a prolonged dry period.
However, it seems that the high capacity of water storage in the sessile oak
vessels in the narrow external rings play an important role in tree survival
during the periods of low soil water availability in the summer. The attenua-
tion coefficient of the y radiation of well-hydrated tissue is twice that in dry
wood and a high correlation between the density of sessile oak sapwood
and its water content was noted by Weibe (1992). It was also observed that
the north-facing quadrant of the trees had a higher density. Tognetti et al.
(1996) confirmed the previous statements of Wiede (1991) for sessile oak
sapwood, namely, that in living trees the density is highly correlated with water
content.
The key point in the interpretation of tomograms is related to the calibra-
tion of the pictures. For this purpose Beres and Fenyvesi (1994) as well as
Raschi et al. (1995), Tognetti et al. (1996) suggested the use of several methods
(ultrasonic acoustic emission, thermoelectric method, etc.) together with a
computed tomography method. For quantitative interpretation of tomograms,
the data deduced with these methods must agree with anatomic studies as well
as with micro densitometric research.
Table 2.12. The influence of ammonia stress pollution of Pinus sylvestris on the sapwood/
hardwood ratio and the moisture content of the trees. (Kiitzel et al. 1997, with permission)
Site Distance Pollution Cross stem Sapwood Hardwood Moisture Content CT number
(m) (area em') (%) Ratio Sapwood Hardwood Ratio
minimum polluted area. The diminishing of the sapwood area induced a reduc-
tion of sap flow in the tree and indirectly a reduction of water supply in the
needles. The increasing heartwood area combined with the diminishing
sapwood zone in the polluted tree can be understood as an accelerating aging
process of the tree. The moisture content of both sapwood and heartwood
zones in the polluted area was much reduced when compared with the unpol-
luted area. The survival prognosis of tree in the polluted zone is reduced. The
results of this research demonstrated that computed tomography with a
portable device is a very useful tool for nondestructive diagnosis of living trees.
During the last 20 years, there has been a growing interest in the development
of automatic methods for quality assessment of logs and lumber and in the
determination of optimal edging and cutting procedures in the lumber indus-
try (Birkeland 1985; Wagner and Taylor 1985).
The first commercial scanner developed in 1973 was relatively slow, requir-
ing 4 min per image. The rotation of a significant mass around a heavy object
like a log limited the rotation speed and increased the resulting scanning time.
Recent technological developments permit the production of appropriate
industrial devices able to detect small structural details. An equipment
resolution of about lOmm (or 1 cm3 in volume) is of interest for the nonde-
structive control procedure. Wagner et al. (1989 a,b) described the first
utilization of an ultra-fast computed tomographic medical scanner with
34 images/s for the detection of defects within hardwood logs. The fast scan-
ning time was achieved by eliminating all moving parts of the device, using a
fan detection system and a reconstruction matrix of 128 x 128 for a water
oak log with a diameter of 38 cm and a length of 3.6 m. The tomograms were
validated by visual comparison with real images of log cross sections, each
8mm thick.
Chiorescu and Gronlund (2000) noted that the recent technological
advancements in X-ray scanning techniques have a good potential for on-line
implementation in softwood sawmills. It has been shown that the sawing
Applications 53
24.00
30.00
36.00
42.00
48.00 --
54.00
60.00
66.00
d)
72.00
78.00
84.00
90.00
96.00
24.00
30 . 00
36.00
42.00
48.00
--
54.00
60.00
66.00
72.00
78.00
84.00
Fig. 2.25. The development of sapwood in Scots pine induced by ammonia pollution, compared
90.00
with a normal tree at 100 em stem height. Trees are located in a forest from the district Torgelow,
Germany (Katzel et al. 1997, with permission). A Tomogram for tree no. 1, with 624-cm' cross
96.00
section located at site 1, 200m from the farm (maximum pollution zone). B Tree no. 8, with 745-
em' cross section located at site 2,280 m from the farm. C Tree no. 2 at site 3, with 513-cm' cross
section located 2900m from the farm (minimum pollution zone). D Tree no. 8 at site 3 located
2900m from the farm (minimum pollution zone). The density scale in CT numbers is represented
by different colors, CT = 30 corresponds to dry conditions and CT = 96 corresponds to a high
moisture content. The tomograms were taken in September 1993 and 1994
54 Ionizing Radiation Computed Tomography
The main interest for the inspection of the internal structure of logs is related
to the improvement of the grade and volume yield in the sawmill. Prior to
studies for the development of commercial industrial tomographic log scan-
ners for sawmill use, feasibility studies, and laboratory investigations were
made. The stress was put on the detection and location of knots, decay zones,
decay areas, cracks, internal holes, abnormal variation of density, abnormal
growth rings, reaction wood, resin pockets, spiral grain, embedded metal, etc.
(Taylor et a1.l984; Funt and Bryant 1987; Swanson and Hailey 1987; Hodges
et al. 1990; Maroc et al. 1996; Niemz et al 1997; Bahyl 1998; Karsulovic et al.
1999; Oja and Temnerud 1999; Nystrom and Earl Kline 2000; Sepulveda and
Gronlund 2000; Varga et al. 2000). The high moisture content in green logs
makes the detection and location of defects especially difficult. Moreover, the
scanning must be performed at a rate compatible with log conveyance speed
in sawmills. During the transportation through the scanner, the logs must
exhibit little side motion, rotation or vibration. The automatic interpretation
of log tomograms started in the 1980s as feasibility studies with automated
defect identification and labeling using different algorithms, first for softwood
and later for hardwood logs. The main objective was to find a procedure
for detection of defects able to operate at real time speed so that scanning,
image reconstruction, interpretation, and display could be integrated into mill
processing.
Scanning applications in sawmilling and wood manufacturing processes, as
suggested by Johansson (1985), can be divided into three main domains:
- the control inspection, where production specimens are checked randomly;
- research and development for new technological procedures;
- on-line internal inspection before sawing at normal production speed.
The following is a review of the literature related to the internal inspection
of logs before sawing. In softwood logs, the knots are the most common grade-
reducing characteristic. As an example, the studies performed by Karsulovic
Applications 55
et al. (1999) are chosen to demonstrate the ability of the linear attenuation
coefficient of gamma radiation to detect the knots in Pinus radiata. The
scanning devise is shown in Fig. 2.26. The irradiation of the samples (40 x 125
x 300mm) was done with a gamma ray beam (24lAm) of 59.5keV, with a 3-mm-
diameter collimator for source and detector. The sample was 131 mm from the
detector, and the distance from the source to the detector was 145 mm.
The array for the irradiation was 5 x 5 mm. The variation of the linear attenu-
ation coefficient in radial and tangential directions of the specimen was mea-
sured. Three zones of gray were used to distinguish knots from the
surrounding zone of defect-free wood. A transition zone is clearly observed
between clear wood and a knot, caused by a strong deviation of fibers from the
natural anisotropic axes.
Using an industrial scanner with automatic image analysis, Taylor et al.
(1984) presented the characteristic image of knots shown in Fig. 2.27. This slice
was compared with the image analysis of the same section and with the pho-
tographic image of the cross section of a southern pine log to determine the
correct location of knots, for which a good agreement was obtained.
In current industrial processing in choosing the best sawing pattern, the
location of knots as well as the log geometry will be determined first. Skatter
(1998) established the cross-sectional shape of softwood logs from three X-ray
projections and modeled the cross section as an arbitrarily oriented ellipse. At
the same time, the exterior shape, the shape under the bark, and the heartwood
zone were determined. The comparison with experimental results on 40 Scots
pine logs gave 1.64% accuracy in the prediction of the model. Secondly, the
best log orientation will be selected to produce the lumber with the highest
value for each log. Funt and Bryant (1987) developed computer algorithms to
detect knots, decay, and cracks in hemlock logs with a medical scanner. The
decay zone was the hardest feature to distinguish and was identified on the
basis of a dark shade and a disruption in growth ring pattern. The program
used a low ring uniformity pattern as a criterion for decay identification. The
algorithm required 3 min/scan, which was very long compared with the speed
of log conveyance in the sawmill. Wagner et al. (1989, a,b) reported the uti-
lization of an ultra-fast medical scanning system for logs of 38cm diameter,
with 34 cross-sectional scans per second. They noted the necessity for the
development of a specific industrial scanner for sawmills. The second-genera-
tion scanners with translate-rotate geometry required 20s/scan. The third and
fourth generation with rotate geometry only required 1 s/scan which is the
maximum speed attainable today in rotating scanners. A short scanning time
can be achieved by eliminating all mechanical movement parts, for example,
by using an X-ray fan scanning system.
Schmoldt et al. (1993), and Schmoldt et al. (2000) emphasized the necessity
for the development of specific computer algorithms for hardwood logs, which
have a more complex internal structure than softwoods. An increase in the
number of log defects increases the difficulty in making the best sawing deci-
sion. For automated interpretation of tomograms, data collection, and image
56 Ionizing Radiation Computed Tomography
a) 0
Fig. 2.26. Sound knots and surrounding area with different levels of gray. a Three levels of gray:
dark zone for the knot, white zone for the sound wood, and gray for the zone around the knot
(Krasulovic et al. 1999, courtesy of Forest Products J). b Tomographic image on a transverse
section of fir, where the knots are visible as well as annual ring pattern, bark, sapwood, com-
pression wood and cracks starting from the core. (Varga 2000, with permission)
Applications 57
........ - ... .
• , ,11'" 1 ....... , ",. • •• "'
I • • • Ut ••••••••• , •• ~ ••••
..............
I ••• U • • • • • • • • • •• ••• • •
~.t
·I·f.~
.re
• It •••••• II • I ••• 1. •• ...
··~········B
"" , , " " 1 " , • •
~ et al. 1984, courtesy of Forest Products J). a
• •• '""';1 . . . . . , • • • • • • "
r;--;- . . Tomographic image. b Photographic image
U •• , •
• t • ~ •• " I • I • I ... II
II •••••••••••••••••••••
(the bolt was split horizontally before sawing
· . . • . . . • . . • " !.. t.. . : . • ~
the disk). c Image analysis
content. The central boards were scanned on all four sides and then graded.
The log classification was considered valid if at least one of the boards was
graded the same as the predicted log grade. The agreement between the
manual lumber grading and the simulated X-ray log scanner was between 71
and 89%. It was concluded that the simulation procedure and grading algo-
rithm was a good predictor for the grades of logs and lumber.
A large scale validation test of computed tomography software against
sawmill cutting was reported by Chiorescu and Gronlund (2000). This test was
based on the Swedish Stem Bank X-ray computed tomography database (Gron-
lund et al. 1994) that acquired information on 600 selected Scots pine saw logs.
The schematic path representation of the validation approach is presented in
Fig. 2.29. The validation approach has two branches, the first corresponding to
the simulated part of the study and the second to an actual medium-sized mill
located in northern Sweden. The conversion process from the sawmill and the
simulation is shown in Table 2.l3. The volume recovery (48.08%) of the sawmill
was very close to the simulation recovery (48.34%). A good prediction of the
quality and length of boards was noted.
Schmoldt (1996) and Schmoldt et al. (1996a) and Schmoldt et al. (1997)
developed specific X-ray computed tomography algorithms for hardwood logs.
Specific scanning parameters are required for X-ray imaging of logs of differ-
real
. ....
i
sim ulated
. . .-. ....
, I
Fig. 2.28. Real and simulated image of a log. (Grundberg and Gronlund 1997, with permission)
Table 2.13. The ability of the simulation software to predict board quality distribution.
(Chiorescu and Gronlund 2000, with permission)
~ ~
[ common sawing process attributes 1
Fig. 2.29. Validation approach of the computed tomography software against sawmill cutting.
(Chiorescu and Gronlund 2000, courtesy of Forest Products J)
ent hardwood species of different sizes, and with different image resolution
and contrast. Data visualization and condensation must form a clear view of
the geometry of the log and of its internal features. Only log defects that are
important for subsequent processing need to be identified. Statistical labeling
of accuracy was introduced, and segmentation and labeling into a single clas-
sification step were possible, largely by using local pixel neighborhoods. A
neural network based on statistical classification of accuracy by using nor-
malized pixel values as inputs to the classifier can formulate and apply aggre-
gate features (average and standard deviation), as well as texture-related
features. The recognition of different hardwood species was based on a
histogram-based preprocessing step that normalized the CT numbers for
density values prior to artificial neural network classification. Morphological
postprocessing is used to refine the shapes of the detected image. The 3D iden-
tification of defects proposed by Schmolt et al. (1997) is a major step in the
development of machine vision for wood species of very complex structure
such as oak and yellow poplar. Initial results using a 2D classifier produced a
lower classification accuracy than 3D data and it was suggested that the 2D
60 Ionizing Radiation Computed Tomography
......-
longitudinal direction
sample 5
1200
1000
a::i 800
u
bb 600
C
.~
'"v~ 400
"0
200
0
0 50 100 150 200 250 300 350
a) longitudinal position (mm)
t
longitudinal direction
line 4
line 3
line 2
line I
sample 5
1200
1000
a::i 800
~ 600
C
.0
.;;; _._. line I
v~ 400
"0 -0 - line 2
heartwood deduced from tomographic inspection of logs, and the stiffness and
strength measured on boards. It was concluded that the mechanical properties
of each board are partly related to the sawing position of the corresponding
board. From the information about the sawing position, it is therefore possible
to predict the strength and stiffness of each board.
In the near future, this important advancement in lumber inspection needs
to be confirmed with more studies with a larger number of specimens of the
same species as well as with other softwood and hardwood species.
Fig.2.30. Density variation observed during length scanning of a southern pine board of 350 mm
length and 38 x 140mm section, containing a knot (Suryatmono et al. 1993). a Longitudinal
position; b transverse position. (Courtesy of Forest Products Society)
64 Ionizing Radiation Computed Tomography
(1992), Pang et al. (1994), Pang (1996), Wiberg (1996), Arfvidsson et al. (1997),
Pang and Wiberg (1998), and Wiberg and Moren (1999).
Kanagawa and Hattori (1985) used a medical scanner for the study of the
distribution of moisture content during conventional kiln drying of 10.5 cm2
lumber of hemlock (Tsuga heterophylla) and of almaciga (Agatis sp.) from
green to 12% moisture content. The moisture distribution on sample sections
was estimated following the procedure specified by Hattori and Kanagawa
(1985) from the CT number of wet wood, oven-dry wood density, and volu-
metric shrinkage. A standard error of 1.1 % was measured for the reference
specimen with moisture content ranging between 2 and 27%. The errors were
caused by the lack of uniformity of the images and lack of perfect linearity
between the oven dry wood density, shrinkage and CT numbers.
The distribution of moisture on radial and tangential sections for both
species are given in Fig. 2.31. In green Agatis sp. (50-70% moisture content) a
high moisture content was observed at the limit of the annual ring and in ray
tissue in the center of the sample. The moisture gradient in the tangential direc-
tion was more important than in the radial direction. In green hemlock
>R
~ 100 100
1::
Q)
1::
0
<.)
a
Q)
en
.~
50 50
o
edge-grain center edge-grain flat-grain center flat-grain
surface position surface surface position surface
36-
36 E 36 E ·85C>950 E .540-590
75<>850
30 _E •.650-150 30 S .490-~
30S "~lIoo
24* :~~
~a_
24 ~ :.~::: 24~ ~= ~ a J40.390
18~ 18 18 5
s:
12; 12-2 12"E
6 6
.
III
6 i S
0
0 10 20 30 40 50 60 70 60 90 100 0 10 20 30 40 50 60 70 80 90 0 10 20 30 40 50 60 70 80 90
d Board wId1h (mm) e Board width (mm) r Board widlh (mm)
36-
E _1()O.l20
36 e :~r:~: 30 E • 110-'00 • 15-18
30.§. .'00.130 - . 60-00 .12-15
• 9-12
24 :ec 24 ~ ~ ~ :::: II 6 -9
~ a 0-20 a 3·6
18~ _18£
12"E = 12l?
6 S 6
.
III
0
0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 0 10 20 30 40 50 60 70
it Board widlh (mm) h Board wId1h (mm) Board wid1h (mm)
Fig.2.32a-i. Dynamics of wood drying (Pang and Wiberg 1998). Experimental tomograms (with permission). a Scanning of the board before drying - corresponding to
wet saturated wood. b Scanning of the board before drying - corresponding to wet wood after 9.6 hours of drying, and to the early stage of the drying process. c Scan-
ning of the board before drying - corresponding to wet wood after 30.4 h of drying and to the late stage of the drying process. Predicted images for density distribution.
d Wet wood density before drying. e Wood density after 9.6h of drying. fWood density after 30.4h of drying. Predicted images for moisture content distribution. g Model
of predicted wet wood density before drying. h Model of predicted wet wood density after 9.6 h of drying. i Model of predicted wood density after 30.4 h of drying
Applications 67
,\.
140
:a...
~ 120 '\.
"-
~ 100 -- -
"
model predicted MC
~- -
-
•
~
.
80 "I> CT scanned MC
-
~
.~
60 •
" ...........
'" 40
~
.......
~
~ 20
--
1- ........ ~
oT
o 5 10 15 20 25 30 35
drying time ( h )
Fig. 2.33. Comparison between moisture content predicted by the model and the experimental
data deduced from the tomograms of Pinus radiata board, expressed by the drying curve. (Pang
and Wiberg 1998, with permission)
The data concerning the utilization of X-ray computed tomography for the
characterization of wood-based composites are very sparse. Sugimori and Lam
(l999) reported the utilization of this technique for the study of macrovoid
distribution in strand-based-wood composites of 630kg/m3 density. The
macrovoids in strand-based composites, of various sizes and random distrib-
ution, are inherent to the forming process of boards. Their presence is due to
the imperfect bonding between the adhesive and the wood strands.
The X-ray computed tomograms were obtained with a medical scanner, set
at 120 k V, 200 rnA, 10-mm slice thickness, and 0.86-mm pixel size and 0.74-mm2
area, in a 512 x 512 pixel matrix. The image reconstruction was done every 3
s, and 130 slices were produced. The size of macrovoids was considered to be
::;;2 pixels, i.e., 1.48 mm2 • The images obtained were converted to black and white
with a conventional image-processing technique, and their size, number, posi-
tion, and major and minor diameters were measured with the medical soft-
ware. The variation along the length of a board of macrovoid area as well as
their distribution is shown in Fig. 2.34. The authors noted that the experimental
size measurements are in agreement with their mathematical prediction.
Gamma ray scanning densitometry (Feraz 1976; Ranta and May 1978) has
been more frequently used for the measurement of the densitometric profile
of wood-based composites. Winstorfer et al. (l986) measured the densitomet-
ric profile of particleboard (l2, 18 and 44mm thickness), waferboard (l3 mm
thickness), fiberboard (6 and 12mm thickness) and medium density fiber-
board {l0, 19mm thickness). A gamma source of 60keV was mounted in a col-
limating holder, and the sample was placed on a carriage that held it stationary
68 Ionizing Radiation Computed Tomography
lowden i~
X"----
400 X direction ( pixel)
o
a)
'(
X=376 X=367
~JI ~
r-~~--------------------------------------~, 1~
.
-
"
t. ..
.. -;.. ..
-- .. .
., -
J
."-." ,.
.....
X ~_~~-L~--~·--LL'~·--~--~--------------"-·---~-·----~o
400 X direction ( pixel )
b)
Fig. 2.34. Macrovoid distribution in particleboard of 630kg/m3 (Sugimori and Lam 1999).
a X-ray image; b the same image after conversion with a 270-threshold value. (Courtesy of the
Japan Wood Research Society)
during counting and then automatically moved to the next step. At each step,
which can vary from 15 Jlm to 30 cm, the counter is automatically read, reset
and restarted. Counts at each step were taken every lOs. The average counting
time for a complete scanning of a board depends on the thickness, and is
between 2 and 5 min. The authors suggested using this as a standard method
for nondestructive evaluation with the accuracy and precision of density pro-
files for wood panel products.
Laufenberg (l986) also studied the use of gamma radiation to measure
the density gradient in reconstituted wood products. The mass attenuation
coefficients were determined for different resins: phenol-formaldehyde
(0.181 cm2/g), urea-formaldehyde (0.182 cm2/g) and isocyanate (0.178 cm2/g) as
well as water (0.196cm 2/g) and wood-red oak and Douglas fir (for both
Applications 69
0.183cm2/g). He noted that the large differences between the mass attenuation
coefficient of resins, wood, and water can induce errors in the prediction of
the wood density of the board. For this reason, the boards must be well con-
ditioned at a minimum 12% moisture content prior to density measurement.
In this case, the expected error of the density prediction is <1 %.
Elbez (1990) reported the utilization of computed tomography with a
medical scanner for the control and characterization of cured adhesives and
glue joints in balsa, spruce, poplar, oak and two tropical species, moabi and ipe.
The adhesives were urea-formaldehyde plasticized with furfurylic alcohol,
polyvinyl acetate with metallic salt and resorcinol phenolic formaldehyde. The
tomograms clearly showed structural details of the glue joints and of the adja-
cent zone composed of adhesive and wood. The density gradient was observed
with a maximum in the center of the joint and with a minimum in the prox-
imity of the wood. The minimum density is probably due to the lower degree
of condensation in urea formaldehyde glue or to the lower polymerization in
the case of the polyvinyl glue. Lindberg et al. (1996), using X-ray CT, measured
water diffusion through acrylate latex paint films with 0.1- and O.4-llm parti-
cle sizes. It was proven that the techniques employed were successful for the
measurement of water diffusion in films. As expected, a O.I-llm particle film
had a much lower water diffusion coefficient than an O.4-llm film.
2.5 Summary
2.6 Annexes
3.1 Introduction
wave is excited on the whole surface of the sample and then monitored, using
both the thermographic technique and lock-in data analysis which very rapidly
provide a phase angle image. This remote method was used for detection of
flaws such as delaminations in wood composites (Wu and Busse 1995, 1996;
Wu et al. 1996, 1997b) because of its sensitivity to subsurface defects and its
imaging capability in short time.
In the case of the so called passive method, the material is subjected to
heating from an external source. The thermal image obtained is the result of
a very complex interaction between the heating source, the material and the
presence or absence of defects. The rate of heat application and the mode
of heating by contact or by radiation are factors of major importance for
obtaining good images with correct resolution. The thermal gradient and
not the temperature itself makes it possible to obtain information about the
integrity of the structure. The thermal gradient is calculated from the temper-
ature distribution. The heat application may be at a single point or over a wide
area. This passive mode produces images that are transient and therefore
difficult to capture for a material that is not subjected to any mechanical stress.
The elimination of the loading fixture makes the rapid testing of objects
of various geometries in situ possible. The time interval of interest for passive
thermal testing is a few seconds. During this time the surface temperature
gradients reach the maximum and yield the greatest resolution of the thermal
image.
An advantage of the passive heating method over the active heating proce-
dure is its ability to produce a temperature distribution without resorting to
mechanical loading of the material. The thermal stress or the thermal shock
is too low to damage the material. The disadvantage of the passive method is
that the thermal images are transient and require a fast recording system to
capture the most interesting images during the test.
Electronic imaging systems or chemical imaging systems can be used for
the visualization of the temperature distribution on surfaces. The most widely
used system is an electronic system with a scanning infrared camera, which
produces an image of an object through electronic detection of infrared radi-
ation emitted from the object.
Chemical imaging is obtained with liquid crystals applied to the surface of
the object. They selectively scatter the light according to their temperature.
The thermal resolution is of 0.007 DC and the spatial resolution of 111m (Wilson
and Charles 1981). A photographic recording system is attached to capture
liquid-crystal images during the test.
Catena (1992) reported the use of an infrared passive technique for the
detection of cavities and for phytopathological diagnosis of ornamental trees.
The infrared camera scanning technique was also used for the detection
of different defects, such as delamination in wood composites (Berglind and
Dillenz 2000) or locating knots and voids in lumber (Tanaka 1994; Quin et al.
1998; Steele et al. 2000a,b; Tanaka and Divos 2000). In wood technology,
the scanning infrared imaging technique seems to be one of the most con-
78 Thermal Imaging
The spatial damped nature of thermal waves has two major advantages over
other tomographic techniques: the high near-surface feature resolution and at
the same time less susceptibility than other waves to the appearance of spuri-
ous signals from a remote interface. For these reasons, tomographic techniques
were developed for clinical applications and for nondestructive evaluation of
critical and valuable manufacturing components.
A laser beam that has a high spectral selectivity, spatial coherence and res-
olution is usually the light source used for generation of thermal waves. The
choice of laser characteristics (wavelength and power) depends on the thermal
properties of the sample. He-Ne laser and Ar+ laser were used for a power level
of about 20 W. They have an acceptable ratio of signal to noise and cause no
optical damage to the sample (Busse 1988). Tsuchikawa et al. (2000) proposed
the utilization of a semiconductor near-infrared laser of 830 nm wavelength
and 48 m W output power for the study of the effect of physical characteristics
of a specimen (sample thickness, fiber direction, surface roughness, moisture
content) and the geometric conditions of illumination on the transmitted
output power.
The relationship between the sample thickness and transmitted output
power for Sitka spruce and beech specimens (Fig. 3.1) shows an exponential
reduction of power for both species. This graph argued in favor of the devel-
opment of a nondestructive technique for measurement of veneer thickness.
The thickness of the sample can be controlled by varying the modulation fre-
quency of the incident radiation and hence the thermal wavelength.
The combined effect of light irradiation and of the anisotropic directions of
wood is shown in Fig. 3.2. It can be noted that the variation of output power
is strictly related to the geometric conditions of the experiment. The effect of
800
incident angle : () =0 0
:i' 700 diving angle : a =90 0
3
~ 600 distance between sample
and detector: dsd =2mm
""
<l)
~ 500
0.
"'[400
g 300
"0
<l)
.§ 200
g'" 100
oI 2
a) b)
Fig.3.1a,b. Transmitted output power under air-dried conditions as a function of sample thick-
ness (Tsuchikawa et al. 2000). a Relationships for beech and Sitka spruce. b Specimen geometry
and parameters. (Courtesy of Forest Products J)
incident angle 0 (deg.) incident angle e (deg.)
moisture content was related to the reduced transmitted output power, which
corresponds to the distance between the sample and the detector of 1 mm for
Sitka spruce and beech specimens of different thicknesses. For both species
and for all thicknesses, the reduced transmitted output power increases with
82 Thermal Imaging
increasing moisture content. This means that the output power increases with
the presence of free water in cell lumina and influences the conditions of light
reflection and scattering.
The second part of this section is devoted to an analysis of some particular
aspects related to thermal conductivity of wood. Because of its anisotropic and
porous nature, the thermal behavior of wood must be related to the parallel
and perpendicular direction of the fibers and to the hygroscopic level, below
and below fiber saturation point. The relationships between the thermal con-
ductivity of wood and some physical parameters such as the density or spe-
cific gravity and the porosity were studied by Siau (1995). The model proposed
by Siau (1995) shown in Fig. 3.3 was used for the derivation of the thermal con-
ductivity equations, for which three elements were considered: the thermal
conductivity of the cell wall substance, of bound water, and of air.
Below the fiber saturation point, when the heat flux is applied in the direc-
tion of the fiber axis, the thermal conductivity is given in Eq. (3.6):
(3.6)
where kL is the thermal conductivity in the longitudinal direction, k Lw is the
thermal conductivity of the cell wall in the longitudinal direction, Va is the
porosity of the wood, ka is the thermal conductivity of the dead air.
When the heat flux is across the cell walls, the transverse thermal con-
ductivity k T can be deduced in a very simplified form by Eq. (3.6) for which
the thermal conductivity of cell-wall substance was 0.318 and 0.401 W/mK for
bound water and 0.024 W/mK for dead air in the cell lumina.
ky = 0.50 - 0.46a (3.7)
-
cell-wall
flux substance
oun water
air
Fig.3.3a,b. Simplified model of a cell wall and relative proportions of cell-wall substances, air
and bound water in wood (Siau 1995). a Geometric model of wood cell. b Relative proportion of
cell-wall substance, bound water and air in wood. (Courtesy of Wood Fiber Science)
Equipment for Thermal Imaging 83
flux ~ lumen
sidewall
The relative flux concentration through the cross walls and side walls is
given in Fig. 3.4. It should be noted that because of the non uniformity of the
flux, the entire width of the wall is not available for conduction.
Above the fiber saturation point, the empirical equation for the determina-
tion of kT as a function of the specific gravity of the wood (G) and of the
moisture content (MC) is given in Eq. (3.8)
kT = G(0.27 + 0.0055 MC) + 0.024 Va (3.8)
The relationship between the longitudinal and transverse thermal conduc-
tivity is given by Eq. (3.9):
kL =(2.25 ... 2.75)kT (3.9)
Thermal properties of wood and wood-based composites have been
reported in several reference books and articles (Weatherwax and Stamm 1947;
Kollmann and Cote 1968; Giordano 1971; USDA Forest Prod. Laboratory 1972;
Siau 1995; Olek et al. 2000) Some values of thermal parameters are given in
Tables 3.1 and 3.2.
In this section, we describe the imaging equipment used for thermal imaging
of solid wood and of wood-based composites. The most popular device for
thermal imaging is the infrared scanning camera, which is a thermal wave
detector in the infrared domain and which can be used for both field and
laboratory measurements. The popular infrared camera can be used for all
thermal wave-type imaging. The thermal wave describes how a temperature
84 Thermal Imaging
Table3.1. Some values of thermal conductivity of wood and other materials in the SI interna-
tional system (Siau 1995).(courtesy of Wood Fiber Science)
Table 3.2. Thermal diffusivity of wood as a function of moisture content and density. (Data from
Kollmann and Cote 1968)
Fig. 3.5. Infrared device for the detection of cavities in trees. (Catena and Catena 2000, with
permission)
86 Thermal Imaging
seen by the infrared camera (Fig. 3.6). The equipment used by Catena and
Catena (2000) for imaging of pathological attacks and internal cavities in trees
is composed of the following subsystems:
- a set of infrared lenses that focus the electromagnetic energy radiated from
the object.
- matrix of 320 x 240 microbolometers pick up the temperature at the tree
surface
- an electronic device that displays the image on the screen
For the inspection of trees, it is possible to use any infrared camera with
high thermal sensitivity and geometric resolution. The images obtained by
Catena and Catena (2000) had a thermal sensitivity of 0.1 °C at ambient tem-
perature and a geometric resolution of 1.3mm at 1-m distance from the tree.
The camera is mounted on a tripod and can be stationed at a variable distance
of up to 20 m from the tree. The average time for the inspection of one tree is
about 2 min. The equipment has the advantage of being non contact and very
easy to use without the need for additional personnel. Table 3.3 gives the char-
acteristics of the lens for the instantaneous field of view of the infrared camera
used by Catena et al. (1990.) for the detection of large internal cavities of trees.
p ---- -
camera-objects
=
I. -
- -==-::.=...-=...-.:.
- ---
distance tEe
F.O.V. (m) 0.13 0.5 0.8 1.4 1.7 2.1 2.5 3.2 3.5
I.F.O.V. 0.13 0.39 0.65 1.04 1.30 1.56 1.95 2.34 2.6
(mxl0- 2)
Fig. 3.6. Geometric description of camera field of view and of the instantaneous field of view of
the sensing element equipped with a 10° X 10° lens. (Catena 1993a, with permission)
Table 3.3. Thermal scanner lens for the instantaneous field of view of sensing elements and the
corresponding field of view of the camera. (Catena et al. 1990, with permission)
Lens Instantaneous field of view The square width side of the field
of scanning element (em) of view of camera (m)
The infrared camera used by Catena (1993a,b) was used in the ambient tem-
perature range from -15 to +55°C. The sensitivity was 0.1 °C for +30 °C object
temperature.
For the nondestructive evaluation of the mechanical behavior of wood
under loading, Luong (1996) used mobile equipment (AGA 782 SW) that com-
prises a set of infrared lenses that focus the electromagnetic energy radiated
by the object, an electro-optical mechanism that determines the field of view
with a scanning rate of 25 fields/s, a set of relay optics containing a selectable
aperture unit and a filter cassette unit, a photo-voltaic shortwave infrared
detector that delivers a signal proportional to the radiation from the object
with a spectral response between 3.5 and 5.6 !lm, a liquid nitrogen Dewar
chamber and a control electronics unit that produces a video signal on the
screen. Figure 3.7 shows the utilization of an infrared camera for laboratory
measurements for thermographic detection of artificial defects and delamina-
tions in plywood. The specimens were illuminated by halogen lamps, and the
temperature was measured in defective and clear zones.
Transient thermography, for which the heat flow is generated by a short light
pulse, is able to give more detailed information about the inspected object, by
a most sophisticated analysis of the temperature differential. Lock-in ther-
mography combines conventional thermography with' modulated thermal
waves. The block diagram of the experimental device for thermal wave imaging
with lock-in thermography is shown in Fig. 3.8.
The sample is exposed to a sinusoidal illumination of a heat beam. A com-
puter controls the sinusoidal illumination power produced by a conventional
halogen lamp. An infrared filter is used to avoid the emission of infrared radi-
ation. During each modulation period, successive scans are performed to
obtain the image from temperature differentials from the magnitude and phase
of the signals. The lock-in thermographic equipment used for the detection of
delaminations in wood-based composites is shown in Fig. 3.9. The phase angle
incandescent lamp
....-------",(1
sample ~
§~E>m'=~'
(mm)
~
420 ~
650
o
color monitor
color
video copy
processor
Fig. 3.7. Infrared camera for laboratory measurements, for thermographic detection of artificial
defects and delaminations in plywood (Xu et a1. 1994; courtesy of Forest Products Society)
88 Thermal Imaging
computer
and
amplifier
sy tern
AGEMA lR camera
sample
(Pc!
~
power supply
Fig. 3.9. Equipment used for the detection of delamination in wood composites samples. (Wu
1994, with permission)
images were taken with 0.03,0.06 and 0.12Hz modulation frequencies. Three
minutes are necessary to obtain an image.
Lastly, we wish to cite one of the most interesting techniques, which is the
lock-in vibro-thermal method (Busse et al. 1992a), for which the thermal waves
are generated by a periodical stress inside the sample and not only at the
surface. The phase of thermal waves with respect to the mechanical excitation
provides information about the integrity of the samples. Figure 3.10 shows
the experimental set-up used by Rantala et al. (1996) for polymer materials.
Applications 89
uitcasonic transducer
and water contamer
JR camera and the control unit
r-. r-I
I ! • ~
o
Fig.3.10a,b. Lock-in vibro-thermography setup. a Amplitude modulated frequency in tension.
b Modulated ultrasonic excitation. (Rantala et al. 1996, with permission)
3.4 Applications
Fig. 3.lla,b. Cavities in standing trees. a photographic image. b thermographic image of cavity
(in blue) in trunk and in branches. (Catena and Catena 2000, with permission)
that pathological attack on the roots of trees can be observed with thermal
imaging. For the nondestructive estimation of the sanitary state of trees in
parks and public gardens, the authors suggested different periods for the
infrared inspections, ranging from 1 year for trees having advanced deteriora-
tion to 3 years for trees with 'medium deterioration' and up to 8 years for
healthy trees.
Thermographic images of timber have been obtained with active and passive
methods. The active methods generate the required temperature field in the
volume of the component under test, which acts as a distributor of the heat
source caused by hysteresis or by thermoelastic effects. One of the major inter-
ests in the development of a thermal imaging technique for the lumber indus-
try is related to the improvement of the accuracy of stress grading. Masuda
and Takahashi (1999,2000) studied the feasibility of this methodology using
western hemlock. Steele et al. (1998, 2000a,b) used passive methods for detec-
92 Thermal Imaging
tion of knots in softwoods and hardwoods. Sadoh and Murata (1993) and
Murata and Sadoh (1994), Tanaka (1994) as well as Naito et al. (2000) reported
the thermographic measurements of the slope of grain in lumber. In this
section, both approaches for timber imaging will be discussed.
For the detection of knots, Masuda and Takahashi (1999) reported measure-
ments on western hemlock lumber under cycling loading in bending and com-
pression. For bending loading, the measurements were performed on lumber
with knots. The specimens had a 45 x 105mm cross section and a length
of 600 mm. The lumber was tested under cycling bending at four points,
starting with 5 Hz. This frequency decreased with increasing bending load
because of experimental equipment limitations. The tension and compression
sides of the specimens were inspected. Two experimental configurations are
discussed, the first one with a specimen including an edge knot at the tension
side and the second one with a specimen including a central knot, for which
the variation of the temperature distribution as a function of loading are
shown. For the bending loading, a temperature rise was observed from 1000 N
loading, corresponding to 120N/cm2• The ultimate failure load was observed
at 2973N, corresponding to 360N/cm 2• Thermographic detection of defects
was started with one third of ultimate strength. Improvement of defect
detection can be obtained with an increase in repetition loading to more than
50 times. The hysteresis effect favored the low thermal conductivity of the
defective zone. Imaging of wood structure and knots was possible under high
speed cycling loading. Clearly defined annual rings were observed at 6000 N
(Fig. 3.12).
I.
18.i
E
~ "
~
~
17,,.
11
16.-5
2000 <000 6000 iIOOIl 10000
a) b) ~1Cdloodt~Gf)
Fig.3.12a,b. Imaging of wood structure and knots at 6000N cyclic loading oflumber. a Image of
annual rings around different knots at 6000N. b L left edge knot; R right edge knot; N central
line of the sample; A, B, C knots in the central part of the sample. (Masuda and Takahashi 1999,
with permission)
Applications 93
nol9
1600
2.5
1400
1200
1000
c
~ 800
-:iI
.Q
600
400
0.5
200
10 15
no 13.
2.5
E
0 1.5
·c
2
"
~
g
0.5 0.5
Fig.3.13a-d. Relationship between temperature rise and energy loss with hysteresis. a Hystere-
sis loop on load-deflection curve. b Static bending. c Static compression. d Dynamic compres-
sion. The number of the specimen is referred to on each curve. (Masuda and Takahashi 1999, with
permission)
94 Thermal Imaging
associated with thermal imaging of the temperature rise induced by the energy
loss in the hysteresis loop during cycling compression loading. Passive
methods for detecting the differential thermal response of knots and of clear
wood were reported by Steele et al. (2000a). The thermographic images were
produced by two methods: by heating with a radio frequency heating system
during lOs at 35kHz, and by heating with infrared quartz lamps of 6000W at
in-line lumber production speed. Temperature differences and differences
between several physical parameters of wood are shown in Tables 3.4 and 3.5.
For softwoods, the difference between the physical parameters of knots and
that of clear wood was more important than in hardwoods.
It was concluded that for softwoods the temperature difference of 2.07°C
between knots and clear wood was significantly large enough to allow the
detection of knots with thermographic imaging methods. For hardwoods, the
Table 3.4. Temperature differences between clear wood and knots in lumber as a function of
species and of the heating system. (Data from Steele et al. 2000a, with permission)
Softwoods
Loblolly pine 3.29 0.63
Eastern white pine 2.10 2.33
White spruce 0.80 1.19
Ponderosa pine 2.66 1.22
Douglas fir 2.07 2.35
Eastern red cedar 1.49 0.65
Hardwoods
Red oak 0.40 0.75
Yellow poplar 0.10 0040
Black walnut 0.19 0.12
Basswood 0.12 0.77
Black cherry 0.56 0.66
White ash 0.49 0.95
Table 3.5. Difference between the mean values of physical parameters of clear wood and of knots
in different softwoods and hardwoods. (Data from Steele et al. 1998, 2000a, with permission)
difference is very small and only 0.31°C, which makes the differentiation
between clear wood and knots quite difficult. Rapid heating with infrared
lamps is less efficient than radio frequency heating for lOs. The on-line valid-
ity of the infrared heating system is strongly dependent on the technological
development of new and more sensitive infrared cameras.
The experimental data reported by Steele et al. (2000a,b) were in agreement
with data obtained by Murata and Sadoh (1994) on heat absorption by knots
of hinoki (Chamaecyparis obtusa) and sugi (Cryptomeria japopnica) lumber.
The rise in temperature of the lumber surface irradiated with incandescent
lamps indicated an increasing heat absorbency by the knots of about 1.6 times
that of the clear surrounding wood.
Quin et al. (1998) employed an infrared camera to determine the influence
of the radiant heating periods and the subsequent cooling times on tempera-
ture differences between knots and clear wood in southern yellow pines. The
study demonstrated that knots can be detected regardless of their diameter in
heated lumber, depending upon the length of heating period and of elapsed
cooling time.
For in situ detection of knots, Tanaka and Divos (2000) proposed measure-
ments of temperature differences between knots, clear wood and decayed
wood in sugi samples of 10 x 10 x 80mm, using the daily natural temperature
variations ranging between 26.3 and 3S.7°C. Figure 3.14 shows the tempera-
ture differences between knots and decayed areas and the temperature of the
air during daytime. The air temperature increases from 5 a.m. to 4 p.m. The
evolution of temperature differences for knots is opposite to that for decayed
areas. A good image resolution was achieved with a temperature difference
of 0.7°C, which implies that the method can be successfully used for the
detection of defects in situ. The detection of knots on wood boards with ultra-
sound thermography was also studied. The amplitude image was obtained with
0.8 r----r=--,
0.6
G '----===::..:.J
t.- 0.4
8
e0.2
~ .,""
~ 0 5 .·. -"'6'''''-.6~:--IT+---rj(-1"""
. 1 ---'1""
2 ---' 14- 1"-5- : Ir hh----.-j
13--'-
a-0.2 clock time (hours) .,i
~ }... 1 j
t
v . ~I~ ~ I~ ~ ~ 1'..
., ' .\.
III! , - \ .. ••
~-0.4 " ' w~,
£ \ l ~ ( .~ .'r{' '.l •
06
-. ~
V ~
-0.8
Fig. 3.14. Detection of knots and surrounding area. Temperature difference of knot and decay
areas compared with the temperature of the air during daytime. (Tanaka and Divos 2000, with
permission)
96 Thermal Imaging
GO.6
o A
.t::
~
6 0.4
o
';>
'is
.g0.2
c:
8
<0
Fig.3.15A,B. Relationship between thermal conductivity, density and grain angle. A Thermal
conductivity and density at constant grain angle. Filled circles The angle of heat flow is 0°; open
circles 15°;filled squares 30°; open squares 45°; filled diamonds 60°; open diamonds 75°;filled tri-
angles 90°. B Thermal conductivity and grain angle at constant density, symbolized as wood with
density ranging between 360 and 680kg/m3. Open circles The density for 360g/m3; open squares
for 500kg/m3; open diamonds 680kg/m 3: open triangeles also for 680kg/m3. (Murata and Sadoh
1994; courtesy of the Japan Wood Research Society)
Applications 9?
and noted that this technique is insensitive to wood moisture content and
temperature.
Naito et al. (2000) verified the feasibility of thermographic imaging, using
the finite element method to demonstrate the transient temperature on the
wood surface. The experimental device is shown in Fig. 3.16. The finite element
is seen in Fig 3.17. The variation of the temperature distribution with heating
time, rising from 10 to 60s is shown in Fig. 3.18. These results were compared
with thermographic images. Good agreement was observed between finite
(deflexion) (load)
····..................................
···· ....
··· ..
L-----I'--,-,
CCD camera;
(uni t: mill)
Fig.3.1? Thermographic measurements of the slope of grain. Finite element model. The heat
flux is represented by the vertical arrows in the central region. (Naito et al. 2000; courtesy of the
Japan Wood Research Society).
98 Thermal Imaging
Fig. 3.18. Temperature distribution as a function of heating and slope of grain deduced with the
finite element method time obtained with the finite element method. (Naito et al. 2000; courtesy
of the Japan Wood Research Society)
elements simulation and the real experimental images. This allows the con-
clusion that the slope of grain can be measured with an infrared camera.
Density variation mapping of the annual ring with high resolution (100 Jlm)
can be achieved with the far-infrared spectroscopy method in the range of T
Hz. The experimental device as described by Koch et al. (1998) is presented in
Fig. 3.19. Density imaging is obtained with coherent far-infrared pulses which
are generated and detected by small photoconductive dipole antennae gated
by ultra-short laser pulses in the near-infrared. The pulses are "guided by off-
axis parabolic mirrors to form an intermediate focus of sub-millimeter diam-
eter through which the wood samples are scanned in a raster pattern. At each
sample point a complete transmitted waveform is acquired and subsequently
Fourier transformed. The transmitted intensity integrated over a given window
is translated into a gray value to finally obtain a pixel. The data acquisition rate
is 20-50 Hz and the time required to generate a two-dimensional image is
typically 10 min, depending on sample size and resolution".
Figure 3.20 shows the density mapping of beech and balsa obtained by the
transmission method through a 1.7-mm-thick microscopic section. In beech,
the dark blue and red areas are zones of low transmission in latewood and in
medullary rays, ranging between 700 and 1400 kg/m3. It is evident that the large
Applications 99
eyeiece
Fig.3.19. Experimental device for T Hz imaging of density variations in the annual ring. (Koch
et al. 1998, courtesy of Wood Science and Technology)
rays have a higher density than the latewood. This is in agreement with data
obtained by Keller and Thiercelin (1975) with an X-ray technique. In early-
wood containing vessels of about 250-l1m diameter, the density is much lower
than in latewood, ranging between 500 and 600kg/m3.
It is well known that in contrast to beech, in balsa only the vessels and fibers
can be identified. In Fig. 3.20, the vessels are very well identified as large yellow
zones. The large dominant red zones are produced by the presence of fibers
with a density of about 470 kg/m\ while the radially oriented blue zones cor-
respond to the parenchyma ray cells with a relatively low density of about
300 kg/m3 and where starch is a major component.
Concluding this study, it can be said that the false color image of density
variation in transverse sections of wood gives us a better understanding of the
anatomical structure of wood.
Since the aim of our analysis is to show the way in which thermal techniques
can illuminate wood structure, it is important to note the pioneering
work of by Moksin (1993), who measured the decay time of opto-thermal
signals observed across a pine sample scanned with a pulsed laser source
with 100-l1m step size. The opto-thermal decay time 't was calculated from
Eq. (3.12)
(3.12)
where a is the absorption coefficient and D is the thermal diffusivity of the
sample.
Figure 3.21 shows the decay curve profile of growth rings pine scanned with
a step size of 100 11m, which is consistent with the earlywood and latewood
zones of the annual ring. The minima represent the thinner and dark zones of
latewood, which are more efficient in heat dissipation than the clear earlywood
100 Thermal Imaging
--
0
Density (g/cm 3 )
2 1.3000 - 104000
1.2000 - 1 .3000
4 1.1000 - 1.2000
-
1.0000 - 1.1000
6 0.9000 - 1.0000
0.8000 - 0.9000
8 0.7000 - 0.8000
0.6000 - 0.7000
10 0.5000 - 0.6000
004000 - 0.5000
12 0.3000 - 0.4000
0.2000 - 0.3000
0.1000 - 0.2000
14
0-0.1000
a 2 4 6 8 10 12 14
a) x (mm)
3.0
Xylem vessels
2.5
Transmission
2.0 _ 567·600
_ 535-567
1.5 502 - 535
_ 470 - 502
1.0 _ 437 - 470
405 - 437
0.5 _ 372-405
_ 340 - 372
0
_ 308-340
0 2 3 4 5
275 - 308
b) x (mm)
Fig. 3.20. Density mapping of a transverse section of two species: a beech and b balsa. The size
of the specimen was 14 x 14 x 1.7mm. (Koch et al. 1998; courtesy of Wood Science Technology)
areas. The distance between two minima corresponds to an annual ring width
of 4.4mm.
For the future, we suggest that this technique be utilized for the mea-
surement of the density profile of wood. This opto-thermal method can be
validated with the X-ray densitometric method.
Applications 101
10
o~!:=====:!Y
o 10 20 30 40 50 60
time / IO-\
Longitudinal 0.41
Radial 0.15
Tangential 0.16
Applications 103
Fig.3.22a,b. Thermal anisotropy of a 3.4-mm softwood veneer sheet at 40% moisture content in
a radial tangential plane; b longitudinal tangential plane. The temperature distribution is dis-
played over a 2 x 3cm area. (Cielo et al. 1988, with permission)
104 Thermal Imaging
:~~'. (b) ,
RodCeruu J'~:: 0.2
10 0.1
80!"..., Spruce Sapwood . 0.8
40 ~ ·0.4
20 (b)~ .0.2
10 ~ '-----.J_-=:::.:,-.. . :.0.1
o 40 80 120 160
percent moisture content
6 -
o LiTdlLiT
5 .. e/ed
-KlKd
4
_A-
A--
(b) ..
.g
e 3 - o=----=======. ____ .__ e-
~:-:::::::--- (a)
2- ~.
01_--1--
o 20
1- - - ' - -
40 60 80 100 120
percent moisture content
Fig.3.24. Relationships between moisture content and the ratios of thermal conductivity; a Filled
circles Thermal effusivity: b filled triangles front surface temperature variation; and c open circles
under laser irradiation for a softwood veneer sample. (Cielo et aI. 1988, with permission)
Fig. 3.25. Thermal image of a veneer sheet of 30 x 30 x OAcm with zones of different tempera-
tures and moisture content. (Cielo et al. 1988, with permission)
lO6 Thermal Imaging
142m
-
I I
I I IW WI I J
I I r-
p
<nfrared heaters C
S
J~ h I ....
I I IW WI I I
I L f-1 L ...J
I
A=3/16 Allen Screw
C= moveable clamp 2.84 ill
p= photocell r-
S= infrared heat sensor p
W=wignut C
S
....
~ h
W W I
r-
p
v-,
S
~
Fig. 3.26. Infrared measuring device for green veneer and lumber S Infrared heat sensor;
P photocell; W wingnut. (Troughton and Clarke 1987; courtesy of Forest Products J)
Thermal wave generation in wood under stress has opened new directions
for studies of wood rupture phenomena. The propagation of thermal waves in
specimens under stress can be monitored by a non contact infrared technique,
Applications 107
OJ
..§ 160i,
G:
"--'l30
~
1;j
[ 120 1
8 I
B I
~II 0 ~
] I
g100 II
~
90 1- - - -
20 40 60 80 100 120 140 160 180 200
c) moisture content %
Okumura et al. (1996) studied the behavior of three tree species under static
unconfined compression. The profile information of the temperature distrib-
ution was related to strain measurements obtained on an array of 66 points
(Fig. 3.29). With specimens under longitudinal compression, a rapid increase
in temperature from 19.6 to 22°C along the fracture plane was observed (Fig.
3.30 at point d). The temperature increase in specimens under radial and tan-
gential compression (Fig. 3.31) was between 20 and 21.2 DC. In both cases, the
highest temperature corresponds to the fracture line. As can be seen, the most
important increase in the temperature was observed in specimens under com-
pression in the longitudinal direction.
Narito et al. (1998) studied the temperature distribution in yellow cedar
specimens under static bending test in four configurations. In this section, we
present the thermographic images obtained for specimens under three-point
bending (Fig. 3.32), with a 5-mm-diameter hole drilled in the compressed zone
of the specimen. At a very low loading level, the temperature increased 0.6°C
under the loading point and the presence on the hole is not observable.
Increase in the loading resulted in a better image of the hole. The rise in
local temperature around the hole is similar to that in the fractured zone of
the specimen.
a) b)
~1,
~I 38
A==;?[12 I 9
c)
: . ·. ~18'30
..................... ~-30
I 87 -100 I
Fig.3.29a-c. Specimens for testing under compression loading. a Longitudinal loading on hinoki
specimen; b radial loading on hinoki specimen; c loading on hinoki, sugi and white seraya
composite specimens. Note that for strain measurements strain gauges were used as a lattice
represented by the lattice dots. (Okumura et al.1996; courtesy of Wood Research Institute, Kyoto
University)
110 Thermal Imaging
..... .
SpecImen hlnold
2 4
a) b) a AmQCJnt of compression (mm)
c)
Fig.3.30a-c. Infrared imaging of specimens under compression loading related to stress distri-
bution in the longitudinal direction. a specimen; b load strain relationship; c thermographic
images corresponding to the points a, b, c, d, e, f from b. (Okumura et al. 1996; courtesy of Wood
Research Institute, Kyoto University)
CH1 .. CH2
. . . .. ..
- - . ~~
~
.. :,... ~-~~.
"E-
oo- ..
CH3 ~
".o·
.......
.. . . ::>
'"
Fig.3.31. Infrared imaging of specimens under compression loading related to the stress distribution in the transverse plane.
Images of specimens at rupture under different compression loads applied in radial (CH4) tangential (CH1) and inclined --
directions vs. the annual rings (CH3 inclined vs. T and CH2 inclined vs. R). (Okumura et al. 1996; courtesy of Wood Research
Institute, Kyoto University)
Hole of 5 mm N
-
diameter
>-3
::r
(l)
(- ,)/: 61J- 20 a
eo.
280
s
~
&.
a) 300
150
'5;100
~
....
~ 50
1/' ~
0
0 5 10 15 20
b) Deflectfon(mm) c)
Fig. 3.32a-c. Infrared imaging of specimen under static bending loading. a Specimen with a central hole under static bending. b Stress strain
curve. c Temperature distribution as a function of loading and strain, corresponding to points 1, 2, 3, and 4. (Naito et a1.1998; courtesy of Wood
Research Institute Kyoto University)
Applications 113
defect. The thermal gradient on the surface is one of the most important para-
meters to by studied, when the temperature distribution is known.
During the last 20 years, a large number of important publications have been
devoted to the detection of delaminations in different materials, such as those
used in the aerospace industry, the composites for a large variety of applica-
tions (Henneke and Jones 1979; Charles and Wilson 1981; Kou et al.1987; Busse
et al. 1992a; Wu 1992; Wu et al. 1992, 1997a; Bauer et al. 1992; Lebowitz et al.
1996; Dillenz et al. 2000) for ceramics (Lyamshev et al.1995), for materials with
micro inclusions (Willis et al. 1997,2001), for electromagnetic materials (Bal-
ageas et al. 1996), etc. The techniques developed for the wood industries are an
extension of technological progress achieved in this field for different wood-
based composite materials (Schulte et al. 1995). The purpose of this section is
to describe the thermographic methods used for imaging of delaminations in
a large variety of wood-based composites such as: glued laminated lumber,
plywood, medium-density fiber boards and chipboards.
f \ I
glued line
t
/ '''\ -*
~------------~~~/-- ~£~\-----
f f
f f
Fig. 3.33. Glued laminated specimen with different zones of discontinuities in glue layer. (Masuda
and Takahashi 2000, with permission)
Fig. 3.34. The thermographic image for a specimen with two 16-cm delamination zones. (Masuda
and Takahashi 2000, with permission)
Table3.7. Mechanical performance of the specimens and the corresponding parameters for
thermally detected defects. (Masuda and Takahashi 2000, with permission)
~_ _."._~Dm
),6
.. 1" h •
•
Fig. 3.35. The geometry of specimens used for the detection of the delaminations between wood
and urea-formaldehyde glue. (Berglind and Dillenz 2000, with permission)
tions, the defect depth and in plane size of a specimen, as shown in Fig. 3.35.
The heat camera operated with infrared radiation between 8 and 12 ~m and
had a temperature resolution of 0.08 K. The heat source was produced by two
halogen lamps of 1000 W each, which were placed l. 7 m in front of the test
piece.
The image contrast was calculated as the difference between the signal para-
meter corresponding to the defect and to the sound area, as can be seen in
Fig. 3.36. The multiple correlation equation (Eq. 3.17) between the contrast,
C, the defect width, D, and the defect depth, L, below the surface layer of the
exotic wood merbau had a coefficient of determination of R2 = 0.90, which
means that the variability of the population not explained by this regression
is only 10%.
C = 1.3+0.28 D-0.91 L (3.17)
The relationship between the contrast and the thermally determined defect
size is shown in Fig. 3.37, in which large defects result in a higher contrast com-
pared to smaller defects, and defects below 3 mm cannot be detected.
Several factors such as inhomogeneous lighting of the specimen surface and
natural variability of wood species, can influence the temperature distribution
and the corresponding thermal image contrast. The effect of wood species on
the detection of subsurface defects of 24-mm width and 2-mm depth is shown
in Fig. 3. 37 in which four species were analyzed namely: oak, merbau, pine and
alder. A higher contrast was observed for merbau, a lower for alder, depending
on the degree of "darkness" of each species. "Dark" species have a high absorp-
tion that gives rise to a greater heat flow that propagates into the specimen,
reflects again the defect, returns to the surface and is detected with a corre-
sponding contrast by the infrared camera. The sources of errors that can influ-
ence the surface temperature of the specimen are: inhomogeneous lighting
of the surface and the natural variability of wood, differences between early-
wood and latewood, zones of vessels and fibers, the presence of knots and of
reaction wood. The penetration of thermal waves was between 2 and 4 mm,
depending of wood species.
116 Thermal Imaging
.. ." "'
Fig.3.36a,b. Infrared image of specimens with defective glue zones and the corresponding signal
as a function of specimen length. a Infrared image where the bright zones corresponds to the
defective areas. The maximum contrast also corresponds to defective zones. b The signal profile
vs. the length of the specimen. (Berglind and Dillenz 2000, with permission)
10
-----
A
9
.,
----
8
::I
c;; ~
> /Z.
..,.y -......
..'""
~
6 ...... Qak
-----
4 -- Alder
/ "'-......
lj\
!!
3
~
U
2
o
10 100
log (Thermal defect size (mm/mm))
Fig. 3.37. Contrast in gray scale between a defective glued area with delaminations and sound
area vs. the logarithm of the defect size for different species. (Berglind and Dillenz 2000, with
permission)
Applications 117
The plywood composed of veneer laminae is probably the most ancient wood
composite known from antiquity. The decorative value of fine veneer is highly
prized. The veneer is produced from solid wood as thin sheets, usually cut on
a rotary lathe. The thickness of veneer sheets that compose the plywood can
be below 1 mm and the plywood thickness ranges between several millimeters
to several centimeters. The laminae are alternated with grain at right angles.
Today, plywood has a wide field of utilization, from furniture to material
for building framing and other structural applications. The development of
synthetic adhesives and the improvement of quality control contributed to the
rapid expansion of the plywood industry, despite the declining size and quality
of the timber resource.
For the inspection of defects in wood composites under a veneer lamina,
Wu and Busse (1995) suggested the utilization of the lock-in thermography
method. The feasibility of this technique was demonstrated, using a sample
composed of several veneer laminae of different thickness, glued on chipboard
or on a solid wood block of a different tree species. The phase images were
taken at 0.03 Hz. The following items were studied:
- the detection of artificially induced defects in the chipboard. through the
surface veneer sheets of various thicknesses. For this purpose the veneer
thickness was varied in five steps from 0.5 to 2.5 mm. At the same time, under
the surface, holes of different diameters were made in the chipboard, as
can be seen in Fig. 3.38. Holes with a diameter greater than 4 mm were
detected.
- the detection of natural defects such as knots present on solid wood blocks
covered with veneer sheets (Fig. 3.39). The detection of knots was possible
until about 2 mm underneath the surface layer.
- the detection of the nature of the solid wood blocks (oak and maple) glued
under the plywood composed of different veneer sheets (Fig. 3.40). The dif-
ferences between the thermal properties of maple and oak can be observed
for all thicknesses of plywood composed of different laminae.
Diamclcr
-'
.
,,'" ...., • ,• •,*"-'\, •, ,,
'.' '-'
~ ~,
I
(mm)
10
I" \ I
~'\
t
.. \ ,,
".
\ 8
-' -' -'
,-, ,", ,., ,.,
'-' '.' '.' . I 6
(; -
'- I
.:' e:, t)
4
:: :: c c ::
a) I b)
,O. mm 2. mm
A
-.-
ISOmm
Fig.3.38a,b. Detection of artificially induced defects in chipboard through the surface veneer
sheets of various thicknesses. a Geometry of the specimen. b Thermal image. (Wu and Busse 1996,
with permission)
.l
veneered wn d
a) b)
Fig.3.39a,b. Detection of knots present in solid wood covered with veneer sheets. a Geometry of
the specimen. b Thermal image. (Wu and Busse 1995, with permission)
nondestructive methods used today for quality control. The quality control
operations among which thermal methods have an important role are:
- drying, for non contact infrared reflectance spectroscopy measurement of
the moisture content of wet and dry chips
- gluing, for the measurement of glue distribution on chips with near infrared
spectroscopy as described by Niemz and Sander (1989)
- panel thickness and its surface properties could be detected by phase image.
The thickness accuracy could be of the order of severaillm. (Wu 1994; Wu
and Busse 1995)
Applications 119
a) b)
Fig.3.40a,b. Detection of blocks of different species under the veneer sheets. a Geometry of the
specimen. b Thermal image. (Wu and Busse 1995, with permission)
Table 3.8. Particleboard production measuring task and corresponding methods used for quality
control. (Data from Schulte et al. 1995, courtesy of Tappi J)
From an engineering point of view, it has been recognized that there is a per-
manent need for the development of different types of lumber joints for frame
systems of large spans and for members of different widths and shapes for
architectural applications. For structural members it is necessary to insure the
quality control of each element, to optimize the design values and to establish
the strength.
Metal plate-connected wood trusses are largely used in agricultural, com-
mercial and residential constructions. Loading transfer from the metal con-
nectors to the wood member is nonlinear and depends also on the grain
orientation of the lumber. When this complex and heterogeneous structural
member is loaded, it deforms as a whole element. Stress concentration occurs,
and plastic deformations, micro cracking, slippage at component interface
and fracture can be observed in wood or in metallic connectors. A better
understanding of the behavior of the joints is necessary for upgrading the
design procedure. Failure modes of the joint include failure of the wood
member within the plated zone, tension or shear in the metallic connector,
teeth pulling out of the wood, buckling of the plate in gaps between wood
members, etc.
Luong (1996) used infrared thermography with an infrared camera to detect
the damage in splice joints under tension and shear loading before failure, as
can be seen in Fig. 3.41. The infrared detector has a response time shorter than
Ills. The temperature patterns are represented by distinct color hues. Each of
them corresponds to 0.2 °C. The energy emitted as infrared radiation is a func-
tion of thermal conduction, thermoelastic coupling effects and intrinsic dissi-
pations generated by viscosity and lor plasticity. Intrinsic dissipation (Luong
1995,1998) seems to be the most accurate indicator of damage to a mechani-
cal, loaded specimen.
The development of glued laminated timber during the 20th century had a
significant effect on wood use and opened new opportunities for creative prod-
ucts. The capability to make engineered structural products of a variety of
forms became possible with finger joints and with relatively small timber of
high quality.
The thermographic detection of defects in finger-joint lumber was studied
by Masuda and Takahashi (2000) with Douglas-fir lumber specimens of
35 x 75-mm cross section and 600-mm length under bending cyclic loading.
The maximum temperature rise was 3°C for a 1.5 J energy loss. In a defective
sample, local variations of temperature are different in tension and compres-
sion zones.
The defective area has high stress concentration in which heat is generated
by the hysteresis loop of the stress-strain curve. The neutral axis zone has a
rather constant temperature, while in the tension area the temperature rises
from 16.75 to 18.5°C at 1000N repeated load. In specimens with better
Summary 121
Fig.3.41a,b. Infrared images of splice joints under shear loading. a Before loading. b Shear stress
before failure. Color hue corresponds to 0.2°C. c Splice joints under tension. (Luong 1996, with
permission)
adhesion, the dispersion of experimental values around the neutral axis is less
important.
3.5 Summary
4.1 Introduction
The microwave scanning technique is a relatively new tool for the characteri-
zation of wood materials. During the last 15 years, considerable progress has
been achieved in the design of scanning probes and in image interpretation
(Martin et al. 1987; Filippini et al. 1990; Chazelas 1991; Yokoyama and
Norimoto 1996; Golosovsky et al. 2000; Kaestner and Baath 2000).
For the correct interpretation of microwave imaging, it is necessary to know
the response of the material to electrical magnetic fields. It is also necessary to
understand the mechanism of contrast and the interaction between the sample
and the probe and to model and to measure the distribution of the electric and
magnetic fields around the probe. Consequently, progress achieved in the basic
knowledge of the dielectric properties of wood and of wood-based compos-
ites will allow new industrial applications of the microwave imaging technique.
Systematic studies on the dielectric properties of different species were
reported from 1948 by Skaar and later by Hearmon and Burcham (1954) and
Brown et al. (1963). The following basic points were discussed:
- the behavior of water and the principal constituents of wood, cellulose and
lignin;
- the determination of the fiber saturation point of wood-water interactions
and the hygroscopic characteristics of wood;
- the measurement and prediction of wood density and moisture content.
The dielectric properties of wood were related to microwave processing. The
advantages of microwave heat processing are numerous, such as: fast process-
ing, small space occupied by the equipment, less temperature degradation of
the products compared with other conventional heating techniques, dimen-
sional stability of the products, selective heating of materials and, lower heat
loss to the surroundings (Dench 1973).
The industrial applications were originally oriented primarily toward
improving of drying and gluing technology. Today, the industrial applications
of the microwave imaging technique are related to the detection of internal
defects such as knots, spiral grain, slope of grain, structural discontinuities, etc.
of logs, lumber, and wood-based composites. Another interesting field of appli-
cation of microwave imaging is the dielectric behavior of vegetative materials,
including leaves, stalks and trunks under various moisture and temperature
126 Microwave Imaging
conditions over a wide range of frequencies (Tan 1981; Ulaby and Jedlicka
1984; Sieber 1985; El-Rayes and Ulaby 1987).
In the first part of this chapter, the dielectric properties of wood are dis-
cussed, including the effects of experimental factors and of physical para-
meters on these constants.
The equipment for laboratory measurements and imaging is presented in
the second part of this chapter. The final part is devoted to the applications for
inspection of forests, logs, lumber and wood-based composites.
The basic aspects of dielectric behavior of wood are treated in several refer-
ence books and articles (James 1989; Skaar 1988; Torgonikov 1990). The dielec-
tric behavior of a solid (Bahr 1982; Zoughi 1990,1996; Ida 1992; Zoughi et al.
1996) is determined by the polarization of atoms and molecules under the
influence of an applied electrical voltage. The polarization can occur at elec-
tronic, atomic, dipole and interfacial levels, and can have different causes that
are strongly dependent on the frequency (Fig. 4.1). Electronic polarization is
determined by the displacement of electrons with respect to the nucleus within
the atoms and occurs in the ultraviolet frequency region. Atomic polarization
is induced by the bending of bonds between the atoms and occurs in the
infrared frequency range. Dipole polarization is caused in wood by the
hydroxyl groups of bound or free water molecules (that are permanent
dipoles), which rotate in the electric field at a relatively low frequency. Inter-
facial polarization is induced by the accumulation of charged ions at interfaces
between different regions within the cell wall of the wood (i.e., between the
middle lamella and the secondary wall, or between crystalline and amorphous
regions).
The mechanism of electrical conduction of wood has been explained by
several theories (Skaar 1988) and it is generally accepted that the electrical
charges are carried by the ions present in the wood cell walls and in cellulose.
In a direct current electric field, cellulose exhibits typical ionic conduction
polarization phenomena. Temperature increase in dielectric solids is produced
by an increase in the energy of the continuous random vibrations of molecules.
When an electric field is applied to a dielectric material, the randomly oriented
dipoles align themselves in a direction opposite to that of the applied external
field. In this configuration, the energy supplied by the field is stored in the mol-
ecules as potential energy. By the removal of the electric field, this potential
energy of ordered dipoles can be converted back into the kinetic energy of dis-
ordered dipoles. The rise in vibrational motions of molecules in a microwave
field is determined by the mechanism of ionic conduction and by dipole rota-
tion, that is, a rotation of the polar molecules under the influence of the exter-
nal electric field. The electric field applied to wood at microwave frequencies
Basic Aspects 127
interfacial
~c
8 ----------
()
.~ dipole rotation
~
:0
------------------
atomic
electronic
frequency
Fig.4.1. Relationship between the frequency and the dielectric constants of solids. (Skaar 1988,
with permission)
['=t:*sin (wt) Fig. 4.2. Relationship between the complex dielectric displacement,
dielectric constants phase angle e and loss tanD angle. (Skaar 1988,
with permission)
D' - - - - - - - - D*
I
I
I
I
[' I
I
I
I
I
I
I
I
I
I
I
L--_.I.-----',---;;..!
[" D"
and
E" = D"/E* (4.2)
where the strength field E = E* sin rot and E* is the peak amplitude of the sinu-
soidally varying electric field. In complex notation, we have:
[E *] = [E'] - j[E"] (4.3)
In other words, in an alternating electric field the amount of energy that
can be stored by wood is expressed by its dielectric constants [E*], which are
tensors and complex constants composed of a real constant [E'] and an imag-
inary constant [E"] and are related to Eq. (4.3). The imaginary constant [E"]
expresses the loss of the kinetic energy that can be dissipated as heat and is
also called the constant of dielectric absorption. The physical meaning of the
imaginary part of the dielectric constant [E"] is related to the attenuation of
the wave and can be correlated with the moisture content (Tiuri et a1.l980).
The ratio of real and imaginary constants E"/E' is called tan () and is currently
used in laboratory measurements.
It is generally accepted that wood is an orthotropic material and the dielec-
tric tensor [Eij] corresponding to this symmetry (Lin 1973) can be written as:
Ell o
[ (4.5)
[E] = ~
o
and
tan Oil o
[tan 0]=
[
~ tan 0 22
o tan~J (4.6)
[E/I]
[E] = E: (4.9)
Table4.1. Dielectric properties of Douglas fir at 8.53 GHz, as a function of structural orientation
and anisotropic directions for different moisture content. (James and Hamill 1965; courtesy of
Forest Products J)
Table 4.2. Dielectric constants at 9.6GHz as a function of wood anisotropy for spruce at 12%
moisture content. (Chazelas 1991, with permission)
Table 4.3. Dielectric complex constants of Douglas fir at 4.8 GHz frequency and 6 and 12% mois-
ture content, calculated with the hypothesis of transverse anisotropy with E22 = E 33 • (James et aI.
1985; courtesy of Forest Products Society)
where Ell is parallel to the fibers and E.l is perpendicular to the fibers. The depo-
larization is expressed by the ratio D = EII/E.l called also "polarization ratio"
and can be calculated using the Stokes parameters defined as:
D = (Q2 + U 2 + V 2)1,I2 I - 1 (4.10)
Table 4.4. The coefficients of dielectric anisotropy kJ> k2' k3 as a function of frequency, at 20°C,
where k, = E'I/E'.c, k2 = tanol/tano.c and k3 = E"I/E".c, for wood density ranging between 200 and 1000
kg/m3. (Torgovnikov 1990)
102 103 104 105 106 107 108 109 10'0 10"
Softwood k, 1.60 1.54 1.50 1.46 1.42 1.38 1.32 1.20 1.16 1.14
k2 1.60 1.70 1.75 1.80 1.90 1.80 1.70 1.50 1.30 1.25
k3 2.6 2.6 2.6 2.6 2.7 2.5 2.2 1.8 1.5 104
Hardwood k, 1.38 1.36 1.32 1.28 1.26 1.22 1.16 1.14 1.12 1.11
k2 1.35 1.50 1.50 1.55 1.75 1.65 1.55 lAO 1.30 1.25
k3 1.90 2.00 2.00 2.00 2.20 2.00 1.80 1.60 1.50 lAO
132 Microwave Imaging
y=a+j~ (4.23)
where a is the attenuation constant (Nepers/m) and ~ is the phase constant
(radians/s). After corresponding calculation, for tan8« 1, it was conclude that:
a=0.5~oRtan 8 (4.24)
and
~ = ~o R. [1 + 0.125 tan 2 8] (4.25)
where ~o is the phase constant of air, defined by
~o = 2TC/A o (4.26)
where AO is the wave length of the electromagnetic wave in air.
In the case of wood material, it has been demonstrated (James 1975; Wert
et al. 1984; James et al. 1985) that £' depends strongly on wood density and
tan 8 on moisture content.
The loss of wave intensity is reflected by the reduction of signal amplitude
and by the phase retardation that are caused by the absorption of wave energy
and by scattering. In addition, the anisotropic structure of wood induces
depolarization of the incident wave. The electric vector of the incident wave is
decomposed into two components in the directions of the maximum and
minimum dielectric constants. When the wave is elliptically polarized, the
minor axis of the ellipse corresponds to the maximum dielectric constant,
which coincides with the directions of the fibers. The polarization is a func-
tion of the thickness of the specimen. Perfect elliptical polarization is obtained
only when the specimen thickness is optimum. For lesser thicknesses the
polarization angle is smaller than the true slope of the grain by a factor cor-
responding to the depolarization index, as can be seen in Fig. 4.3.
The effect of density on the phase and on the attenuation of waves was
demonstrated by Kharadly (1985) for hemlock (Fig. 4.4) at 11-12% moisture
content. The corresponding regression equations are given in Table 4.5. For
both parallel and perpendicular polarization, an increase in the density
increased both the attenuation and phase shift. These experimental data were
used for the abacus presented in Fig. 4.5, from which it is possible to determine
both wood density and the moisture content of a board by measuring the atten-
uation and the phase shift of a microwave regression equation (r = 0.87) of the
phase shift for different species as a function of the density, ranging from 400
to 800 kg/m3 (Fig. 4.6) at 9% moisture content as was reported also by Martin
et al. (1987). The effect of moisture content on the attenuation, phase angle and
depolarization index was studied by James et al. (1985) with Douglas fir as can
be seen in Figs. 4.7, 4.8 and 4.9. Increasing moisture content caused an increase
in all these parameters.
Values of attenuation coefficients for Indian wood species with nearly 9%
moisture content, as reported by Jain and Sanyal (1996), are given in Table 4.6.
The measurements were performed at 8.2 GHz in the X band microwave region
134 Microwave Imaging
/ I
X / /
/ / / / grain direction
/ ~
I /
/ /
z
I
a) -t-
X' X
-.....
/ e
I ~
'"
L.h.pol.f
\ \ ,~
E inst
y , ~ain direction
b) \
Fig. 4.3. Slope of grain and the elliptic locus of the instantaneous electric field. a Grain direction
at angle 0 in the plane Y -Z in a specimen of thickness t. b Elliptical polarization of transmitted
instantaneous field. The ellipse is inclined at angle e; the sense of rotation depends upon whether
the X-field component leads or lags the Y-field component in the time phase. The locus becomes
a circle when the two components coincide grain direction and instantaneous E inst. (King 1978;
courtesy of Forest Products Society)
In the first zone, the attenuation increases very slowly and is probably due to
the macromolecular water being very strongly bonded to the wood structure.
In the second zone, corresponding to the presence of "bound water", the atten-
uation dramatically increases with moisture content. Over 30% "free water" is
Table 4.5. Relationships between the variation of amplitude (~A) and of phase (~cp) of the
microwave signals and density (p) of hemlock wood samples at 20°C and with the moisture
content M, constant. (Kharadly 1985; courtesy of Forest Products Society)
parallel polarization
HEMLOK (# 201 300)
MOIST. CONT.=11-12%
+ NORMALIZED PHASE SHIFT +
-..•
+
«+
XNORMALIZED ATTENUATION ..... +
APHASE SHIFT / ATTENUAT~"
0=0 . ~~
;= 540,7 P + 38,0 x
f >s<~x ~'S<
• x~~)( XX
i{Jx
+ )t~x
x ~
X
c
c4---~--·--,-----;----r----;----r----.----r----.----+1
0,0 0,2 0,4 0,6 0,8 1,0
a) density (glem 3 )
Fig.4.4a,b. Attenuation and phase shift as a function of density. (Kharadly 1985; courtesy of
Forest Products Society)
136 Microwave Imaging
+
0 perpendicular polarization
-
0 0
0" 0 .+ 0
....
('.l 0 ('.l
M
+
+NORMALIZED PHASE SHIFT •
-
XNORMALIZED ATTENUATIO~
~o" 0
--:--0. .
~HASE SHIFT / A TTENUATI~ +
o '""'
/
~g 1C5~
0"""
~('.l '-'
~- • -90
Cl
~
; - 497,8p + -25,5 ~
-~ -
... f::
[-<
[-<0 ::C: o
0 ++ 0 ~
sf [-<a'i
00
}i.+
<0 ~oo
~oo
00 -
a'i
[-< ~
[-<
<
[-< ~
.".. xxX Cl
< Cl ~
~o
---
[-< 0" ~o + ~ 0
c.o
~
....:I
~~
... 0
<
EeI.O
00
~
X ~
0
00 0 Z
Z
~o
~o
0
0" 0
"""
1.0
X """"
•
0
0
0 0
0
('.l 0"
0,0 0,2 0,4 0,6 0,8 1,0
b) density (glcm 3 )
Fig.4.4a,b. Continued.
Table 4.6. Attenuation of microwaves at 8.2GHz frequency measured on several Indian species.
(Jain and Sanyall996, with permission)
"The attenuation constant was calculated with the relationship; a = 2.303 X2-'. (lOg1O F2-log1O F,)
(4.27), where F2 and F, are the corresponding intensities of microwaves for wood samples. The
measurements were performed in the X-band region of microwaves on the E10 mode of vibration.
Basic Aspects 137
¢:::
:.a PI
VJ
C)
VJ
o:l
..c
P2
0-
"0
C)
<-<
;:::I
VJ
o:l
C)
E P3
P = density
P4 M = moisture content
constant M lines
measured attenuation
Fig. 4.5. Abacus for the determination of unknown density and moisture content as a function
of variation of phase and amplitude of the microwave signal for hemlock wood samples at 20 0 •
(Kharadly 1985; courtesy of Forest Products Society)
present, and the slope of the regression relating the attenuation and moisture
content is less steep than in the second zone. The fiber saturation point was
found to be between 26.9 and 28.2%, as the intersection between the regres-
sion lines in the second and in the third zones.
The basic physical parameters of wood related to the detection of defects with
a microwave technique are: the moisture content, the density and the grain
angle (the slope of grain). These three parameters can be determined from the
dielectric tensor. The relationships between the diagonal terms of the dielectric
tensor and the moisture content, density, grain angle and chemical constituents
138 Microwave Imaging
o
o
0\
o
o
00
0
0
f"-
6-
<t::
:E
'" °0\0
II)
gj
..<:: c
~
<., c
0
0
c p=107 b=-1,96° r=0,87
11"\ c
c
°0 c
"""
0
0
""0,3 0,4 0,5 0,6 0,7 0,8 0,9 3 3
X 10 kg/m
Density
Fig.4.6. Relationship between the phase shift and the density of different wood species,
with density ranging from 400 to 800kglm3 • (Martin et al.1987; courtesy of Wood Science and
Technology)
0
N
e )
--~
0
......
11"\
~
t;
I':
0
0
0
I': ......
0
.~
6
~
o Parallel
• Perpendicular
Fig.4.7. Attenuation of the signal as
I I I a function of moisture content in
50 100 150 two anisotropic directions, parallel
and perpendicular to fibers. (James
et al. 1985; courtesy of Forest Research
Moisture content (%) Society)
Basic Aspects l39
~
§
t'i
.: 0
0
u 0
11)
~
0;
..<:
0...
o Parallel
• Perpendicular
I , , , , I I , ,
50 100 150
Moisture content (%)
0
0
.-
5",
~t--
;.<
11)
"0
.5
':0
.9 '"
(;j
N
'ta
"0
0..
11)
"0",
.--=: N
.:
=:J
Fig. 4.9. Depolarization index as a
function of moisture content in two
0
anisotropic directions, parallel and
a 50 100 150 perpendicular to the fibers. (James et al.
Moisture content (%) 1985; courtesy of Forest Products Society)
,
__+__
I
I
._._~_-G
00
6
t-
O' .,.,'"
§
<:!
<Il
0..
-
"1
'T
,
e 0()
""II
0
<>
e
.Sl
'"o· 'T'
ta <:!
;:l
is
~
.,.,
6
<')
Europe, USA and Japan, and also from the tropical zones in India and Malaysia
(Hearmon and Burcham 1954; Uyemura 1960; James and Hamill 1965; Lin
1967, 1973; Norimoto 1976, James 1975; Norimoto et al. 1978; Peyskens et al.
1984, Chazelas 1991; Choffel et al. 1992; Goy et al. 1992; Jain and Sanya11996;
Obataya et al. 1996; Dubey and Deorani 1997; Kabir et al. 1998; Greaves 1998;
Zielonka and Dolowy 1998).
For brevity and simplicity, from this important literature related to the
dielectric constants of wood, the discussion will be limited to considering the
relationship of the dielectric constant to the moisture content, to the structural
orientation that induces anisotropy and to wood density.
Basic Aspects 141
0,36 18
0,32
• 1.
16
/ ~ ~In
.....
0,28 / ~ll 1i". V 14
V ~ V W
I ~ /'
0,12 / c~~ V -, ~ 6
• l} _u<, ~
.,co~": ~
0,08
,
¥
~~
/
C1"
l..cf"
/
;r
~pV ._\, ~
~
4
o
5 10 15 20 25 30 35 40 45 50 55 60 65 70 75
Fig.4.11. Dielectric constant £', and tan/) of Douglas fir as a function of the moisture content.
(James and Hamil11965; courtesy of Forest Products J)
142 Microwave Imaging
3
LT
. LR RT
/
....... ---~O
/
/
/
0/
" "- "-
2 r '0
' -
w y-....
........ 0/ .....
.....-C - - - - 0 - ---0--_...,
T J J L L I R R i 1 T
90 60 30 o 30 60 90 60 30 o
grain angle ( deg ) ring angle ( deg )
Fig.4.12. Variation of the dielectric constant in three anisotropic planes. (Norimoto and Yamada
1972, with permission)
The effect of anisotropy can be expressed as a ratio between the dielectric con-
stants, as demonstrated by Makoviny (1988) for two frequencies and four
species: spruce, pine, beech and oak (Table 4.7). For all species it was noted
that E'L > E'R > E'T. In the transverse plane, the ratio E'R/E'T ranges from 1.02 and
1.08 for softwoods and between 1.13 and 1.16 for hardwoods. This means that
from a dielectric point of view the softwood species are less anisotropic than
Basic Aspects 143
6
2'
L,R, T ;
L
5
4 T
R
3
m.e. (%)
1+-----,------r----~----~~----r_----~-
a) 5 10 15 20 25 30 35
2"
2.0 L,R,T:
L
1.5
.. , T
1.0
•. ' R
0.5
m.e. (%)
O~O •• -
b) 5 10 15 20 25 30 35
tan 8
0.4
L, R,T ;
0.3
'" L
.. ' T
R
0.2
0.1
..'
..' m.e. (%)
O.O~----,_----_r----~----_,------r_----~-
C) 5 10 15 20 25 30 35
Fig. 4.13a-c. Dielectric constants as a function of moisture content and anisotropic direction in
pine. a E' as a function of moisture content; b En as a function of moisture content; c tano as a
function of moisture content. (Peyskens et al. 1984; courtesy of Wood Science and Technology)
144 Microwave Imaging
Table 4.8. Attenuation coefficients of microwaves at 10 GHz as a function of sample thickness and
of the anisotropic direction. (Chazelas 1991, with permission)
• To avoid the effect of wood structure on the measurements, the thickness of the sample was
progressively reduced.
a) O ~~---r--~---'
0.1
-50 0.Q75
G -100 l--_0.0_5~
~
~ O%RH 1J %RI-J
.g~ -I 500 ~==:::;::===;;::=:::;;:==~
r-
0.1 0.1
Oo
E 0.2
B
-50
0.15
33%RH 43%RH
-\50 + - - - - - - - -- ---'
b) 0-
-50
85 %RH
- 150 3 4 5 6 3
logarithm of frequency
Fig.4.14a,b. Iso-curves of the dielectric loss as a function of temperature and frequency for the
relative humidity, ranging from 0 to 94%, calculated for hinoki. (Yokoyama et al. 2000a; courtesy
of the Japan Wood Research Society)
which the wood samples were tested at constant temperature (Fig. 4.14). The
distribution of relaxation times narrowed with increasing relative humidity.
was studied by Nishino and Norimoto (1990). The theoretical development was
based on models in which the proportion of chemical constituents of wood
was deduced from the rule of mixtures. The dielectric constants of the princi-
pal constituents are given in Table 4.9. These values were used for estimating
the dielectric constants of the cell wall shown in Table 4.10. Closer examina-
tion of these data and of a simplified model of the anatomic structure of hard-
wood leads to the theoretical values of the dielectric constant £' as a function
of moisture content in the range 0-15%, for three anisotropic directions. The
estimated values in the longitudinal direction were higher than those in the
radial and tangential directions and were in good agreement with data
measured with mizume (Betula grossa).
Table 4.9. Theoretical values of dielectric constants of chemical constituents of wood. (Nishino
and Norimoto 1990, with permission)
0 5.60 4.68
10 6.42 5.84
20 7.54 6.72
30 8.62 7.54
40 9.90 8.49
50 11.58 9.72
60 13.34 10.97
70 15.65 12.60
75 16.87 13.64
148 Microwave Imaging
b) incident field
surface probing
slice 2
Fig.4.1 Sa,b. Microwave imaging of an object. a With projection produced by direct measure-
ments; b with tomography obtained by numerical processing of the wave front. (Bolomey and
Pichot 1990, with permission)
Equipment for Dielectric Measurements and for Microwave Imaging Technique 149
to planar objects, for which the distance of observation must be within one
wavelength.
In contrast to projection imaging, tomographic imaging is obtained with a
more elaborate processing of a microwave wave front and gives a cross-
sectional representation of the object (Fig. 4.15b). Either a space domain iter-
ative approach or a spectral formulation for diffraction tomography can be
used. In the latter case, using the Born approximation, the image is obtained
after taking the inverse Fourier transform. For inhomogeneous solids, the pro-
cedure of reconstruction is subjected to various artifacts due to diffraction
effects. The most promising approach for microwave imaging is the space
domain iterative formulation, which is very sensitive to the accuracy of local
measurements. The profile of the object is obtained from the complex permit-
tivity values. The domain of validity of this method depends directly on the
signal-to-noise ratio. For industrial applications, cost considerations are of
great importance and a compromise between the cost of the equipment and
the rapidity of the performance must be found.
In the following, we will analyze first the equipment for laboratory
measurements of dielectric constants of wood and, second, the equipment for
imaging the structure of wood during processing under industrial conditions.
orthomode
transducer is
elliptically Ol
linearly OQ
polarized polarized S·
OQ
incident transmitted
a) wave (45°) wave
wood specimen
I microwave ~ orthomode
source transducer
linearly ~ elliptically
polarized polarized
incident transmitted
b) wave wave
cross-polar branch
Fig.4.16a,b. Microwave systems for the location and detection of different parameters of wood (Kharadly 1985). a System for the
measurement of attenuation and phase shift for detection of wood density and moisture. b System for measurements of polariza-
tion of transmitted waves for the detection of the slope of grain.
Equipment for Dielectric Measurements and for Microwave Imaging Technique 151
wood specimen
l;n~c1y 'V\J'-.-~"VIf'-'
polarized
lin=ly
polarized
incident transmitted
wave wave
(c)
(d) (a) strip line (b) split waveguide (c) free wave
Fig. 4. 16c,d. c System for the detection of the discontinuities introduced by the presence of knots
(courtesy of Forest Research Society). d Sensors for transmission measurements (Thompson
1996, with permission)
and with a sample, by adjusting the variable attenuator and phase shifter.
This system was designed to measure the moisture content and density of
wood.
The slope of grain measurements are based on the measurements of the
variation of the polarization of a transmitted wave that is elliptically polarized
by the wood structure (Fig. 4.16b). The orthomode transducer separates the
components of the transmitted wave into a co-polar component and a cross-
polar component, which are alternately sampled with a high-speed microwave
switch before processing of the corresponding signal by the detector.
For the detection of defects, such as knots, both the variation of density and
the slope of grain must be measured. For this purpose, a combination of the
previous systems can be used, as shown in Fig. 4.16c. The incident wave is
polarized at 45° versus the direction of the fibers, which generates an ellipti-
cally polarized wave with two components (the co-polar and the cross-polar),
which are detected separately. In this way, a typical response signature is
obtained, indicating the presence of the defect.
Based on a similar approach, Forrer and Funck (1998) used a dielectric spec-
trometer operating in a frequency domain at 10 MHz, for detection and 3-D
imaging of knots. The electrode of the spectrometer was 19.5 mm in diameter
and was mounted on a translation table. The scanning line was chosen to
152 Microwave Imaging
~
...... 0,06 ..Q
V>
"Q
N 0,04
... dielectric constant ... loss tangent
0,02
0 0
0 I 2 3 4 5 6 7 8
distance along scan li ne (em)
Fig. 4.17. Signature of a loose knot expressed by the dielectric constants and by the loss tangent.
Above The location of the scanning line on a loose knot. Below Dielectric constants as a function
of the position on the sample. (Forrer and Funck 1998; courtesy of Holz als Roh und Werkst.)
~o
=-~
f~\0~'
~/(~,~~)'\~\
'''~
~:~~~~*\~\""
~d ' '; ;\ ~
~ ~~) ,
I ~ \ FigA.18. Three-dimensional images of a
tight knot 14mm in diameter. (Forrer and
Funck 1998; courtesy of Holz als Roh- und
Werkst.)
cover a zone of approximately 5 cm 2 in the grain direction with the electric field
applied through the specimen in the radial and tangential directions as can be
seen in Fig. 4.17. The location of the loose knot is precisely determined by both
the dielectric constant and the loss tangent. A 3-D image of a tight knot of
14mm diameter is shown with good resolution in Fig. 4.18.
The limitation of microwave through-transmission or reflection methods is
caused by the low sensitivity of this technique to differentiate between differ-
ent types of defects, as can be seen in Fig. 4.19. Forrer and Funck (1998) studied
the dielectric constants of knots, pitch pockets, clear wood, blue stain and open
holes as functions of moisture content and frequencies (Fig. 4.20). From this
Equipment for Dielectric Measurements and for Microwave Imaging Technique 153
grey levels
255
threshold
\
Fig.4.19. Microwave signals for different types of defects. (ChoffeI1999, with permission)
graph, it was impossible to define precisely a typical defect, but a general sep-
aration in three main groups (air gaps, pitch pockets and tight knots) occurred
with increasing moisture content and frequency. In a large cloud cluster of
data, open holes and loose knots are in opposite position to tight knots.
Between them, clear wood, blue stain, and pitch pockets can be observed.
Forrer and Funck (1998) noted that under laboratory conditions, the main
sources of errors associated with dielectric measurements of wood containing
defects is induced by the moisture content variation in the defect zone com-
pared with the area of surrounding sound wood and by the variability of the
signal introduced by the wood sensor interface.
a) S! • knot
c) s:
- n;lrh ch
o :1""" 'ood
, I ilch lrea.k
pen Olcs
• blue . tain
,.
7-
~~:
. 11
l
,
'1'/
i r
~~
" J~
,v:.
if.:::l>
~
o O.os 0. 10 0, 15 0)0 0.25 0.30 o 0,05 0. 10 0.15 0.20 0,25 0,30 o 0,05 0.10 0,15 0.20 0,25 0.30
a loss tangent b loss tangent c loss tangent
d)=: • kno~
-
o clear ood
I ilch streal -
••
1 ppen o les
• blue lain
;; ""
,I' ,0' ,I' ~
~~
lI!J I~ ,\I'!J
.~" r+
,
M 25% t~r+ MC=2S% .1It· MC=2S%
f,!l f~ L4MHz _ I(IJ~ 10MHZ ~N
f=20MH2
N
r;i
N
1.)
N
ri I
:i! ,< '"
o 0,05 0, I0 0,15 0,20 0.25 0.30 h 0 0.05 0, 10 0, I5 0,20 0.25 0.30 0 0.05 0, I0 0.15 0,20 0.25 0.30
g loss langenl loss langen. loss langenl
g)=: I knOI
h) 0
-
o clear 'ood
I pitch Irca
1 open o les
• blue lain
~ I
~ Jt ~.
II
.2iI
1i~ ~~
,~ . ~
~~
' )Ii
i MC= IO% ;~ MC=IO%
_11 .~~A
MC= IO%
i' f=20MHZ
1-.,.. \.
f=1.4MHz Jfl:( : f= IOMHZ F'
:l1'li '
o 0,05 0, 10 0. 15 0,20 0.25 0.30 o 0.05 0.10 0, 15 0.20 0,2 5 0,30 o 0,05 0, I0 0,15 0,20 0.25 0,30
d los tangent e loss tangent f loss tangent
Equipment for Dielectric Measurements and for Microwave Imaging Technique 155
or in the near field. For far-field imaging the smallest detectable defect is deter-
mined by the ratio d/A >1, where d is the length of the defect in the plane
normal to the microwave propagating vector and A is the microwave wave-
length. This condition is known as the Rayleigh criterion, which is applicable
to all electromagnetic imaging in the far field. For near-field imaging, the
resolution can be as high as A/IOOO, because in this case Rayleigh's criterion
does not apply.
The probes used to evaluate the properties of the samples can be scat-
terometers or radars that measure the scattering properties of the medium, or
can be reflectometers that measure the reflectivity of the sample due to its
inhomogeneities and defects. The probes can operate in the proximity of the
medium as open-ended coaxial lines, cavity resonators, antennas, etc. By
moving the position of the transmitter across the sample, an image of the
structure can be obtained. The development of microwave systems for online
measurements of the moisture content in wood and wood-based composites
is directly related to the economic benefit in the sawmill and particle board
industries for front-end processing, in-process control and output quality
control.
For online wood processing, the microwave equipment must satisfy several
criteria, including:
- the recognition, location and mapping of defects, such as knots, reaction
wood, slope of grain and other major discontinuities of wood structure;
- the capability of operating over a wide range of temperature, wood mois-
ture content and density of different species, with sharp resolution;
- the normal speed of operation required by the processing;
- a non contact and nonhazardous technique, with no susceptibility to vibra-
tions and with simplicity of operation;
- inexpensive training for the operators.
The actual microwave equipment for the inspection of logs, lumber and
composite boards is discussed below.
The inspection of logs is performed with the experimental device
(Kaestner and Baath 2000) shown in Fig. 4.21. In front of a wide-band antenna,
the log is rotated and translated. The frequency band used for measurements
was between 4 and 8 GHz. More details about the experimental results obtained
Fig.4.20a-i. Relationships between the dielectric constant and loss tangent for knots, pitch
pockets, clearwood, pitch streaks, open holes and blue stain at moisture contents ranging from
6.6 to 25% and for frequencies ranging from 1.4 to 20 MHz. a Moisture content 6% and frequency
1.4 MHz. b Moisture content 6% and frequency 10 MHz. c Moisture content 6% and frequency
20 MHz. d Moisture content 10% and frequency I.4MHz. e Moisture content 10% and frequency
10 MHz. f Moisture content 10% and frequency 20 MHz. g Moisture content 25% and frequency
1.4 MHz. h Moisture content 25% and frequency 10MHz. i Moisture content 25 and frequency
20 MHz. (Forrer and Funck 1998; courtesy of Holz als Roh- und Werkst.)
156 Microwave Imaging
----'x
n knot
b)
Fig.4.21a,b. Scanning oflogs with a microwave device. a Components of the device. b Refraction
indices at the interface between log and knot. I10g and Ikno, are the distances from the antenna to
the log and the knot; nai, and nkno' are the refraction indices at the interface. (Kaestner and Baath
2000; courtesy of Forest Research Society)
with this equipment are given in Section 4.4.2. Technical details concerning the
equipment used for the measurement of phase and amplitude for the inspec-
tion of lumber are given by King (l978), King and Yen (l981) and Martin et al.
(l987).
While microwave techniques have shown a great potential for inspection of
lumber, their applicability seems to be limited by their inability to recognize
the nature of the defects. To overcome this drawback (Choffel and Martin 1966;
Choffel 1999), it was proposed to supplement the equipment with a CCD
camera, thus providing a coupling of a vision and a microwave device (Fig.
4.22). In this system, wood structural data are correlated with lumber strength
and compared with classification grades. The assignment of a grade to a piece
of lumber was based on the detection and location of defects and on three
physical and mechanical parameters: the density, the strength and the stiffness.
The camera provides the profile to calculate the dimensions of the board and
to detect the presence of defects, while the microwave sensor shows the inter-
nal structure of the wood. The network is described by Beatty et al. (l999) and
is based on the identification of the characteristic microwave signature of
lumber that distinguishes one piece from another. The accuracy of the in-
spection of the internal structure depends on lumber thickness, operating
frequency and the input power level.
The system described here has some advantages, such as: relatively low cost,
satisfactory design complexity, good spatial and measurement resolution, no
extensive need for signal processing and no radiation hazard. The system
requires no contact and has a rapid response to automatic process control.
Equipment for Dielectric Measurements and for Microwave Imaging Technique 157
lights
Linear Camera
1728 pixels
256 Grey levels
Knots sensor
Halogens
Conveyor
b)
Fig.4.22a,b. A vision and microwave-coupled device for detection of defects in lumber. a The
device used for lumber inspection. b Details of the system of microwave probe and vision camera.
(ChoffelI999, with permission)
md}
m", m
ref. ....
rp 8 me
phase and
amplitude A
e-< :>
0-
E p", }p
Pd
0
detectors u
e'
e" }e
Fig. 4.23. System for inspection of composite boards with a homo dyne instrumentation. (King
and Bausel 1993; courtesy of Forest Products J)
4.4 Applications
This section is intended to cover a wide range of topics related to the applica-
tion of microwave imaging techniques for the inspection of forests, logs,
lumber and wood based composites. The microwave parameters used for
image reconstruction are: the amplitude, the phase and the polarization. The
system involves transmission of microwave energy through the samples or
objects and measurements of reflected, refracted or diffracted fields. Two- or
Applications 159
00 balsam fir
trunk
Mv=0,166
.." .. .
...
t'11
o Co o c
o
o 2 4 6 8 10 12 14 16 18 20 22
frequency (GHz)
Fig.4.24. Dielectric spectrum for f' and fN of a balsam tree trunk in a frequency range from 1
to 22GHz. (EI-Rayes and Ulaby 1987; copyright 2001 IEEE)
5,5
x(cm)
Fig. 4.25. Radial variation from pith to bark of the complex dielectric constant f*, of a fresh cut
balsam fir tree. (El-Rayes and Ulaby 1987; copyright 2001 IEEE)
demonstrated that the estimation of the woody plant biomass and the
monitoring of different temporally dynamic processes are possible by using
spaceborne synthetic aperture radars (SAR). The best performance for
biomass estimation was achieved with the lower frequency (P- and L-bands)
radar systems.
The penetration measurements at the L-band for coniferous and deciduous
trees are given in Fig. 4.26, for horizontal (HH) and vertical (VV) polarization
directions using the experimental arrangement shown in Fig. 4.27. The data
were collected at I-m distance along blocks of ten trees. The mean backscat-
tering cross section was calculated as a function of the height, ranging from
the ground to 6m. For coniferous trees (spruce) the backscattering between 1.5
and 4m is relatively constant at -20dB for both HH and VV channels and
increases dramatically between 4 and 6m. The attenuation is higher for hori-
zontally (HH) polarized L-band waves than for vertically polarized waves
(VV). The behavior of hardwood trees is different, and the hypothesis that the
radar acts as a shape filter was advanced by the authors. They proposed that,
for hardwood trees, the volume of branches is large enough to act as prime
scatterers. For coniferous trees, however, the tree area projected in the direc-
tion of the incident radar wave seems to be the most important parameter. The
backscattering characteristics of spruce trees were investigated as a function
of polarization angle, ranging from 0° to 100°. At 0° the electrical field vector
was oriented parallel to the trunk. The signal obtained was different, depend-
ing on whether or not the trunk was illuminated (Fig. 4.28). The stem strongly
scatters the incident wave, and the increasing polarization can be related to the
decrease of its cross section with height. The slope of the curve between 0° and
35° can be related to the diameter of the trunk. At a 43° polarization angle, a
maximum of the curve was observed. This angle was presumed to correspond
to the preferred orientation of the major branches. The peak at 43° was also
observed for a tree whose trunk was not illuminated. For polarization angles
larger than 43°, both curves (HH and VV) have the same appearance. This
aspect was related to the high probability that other branches were oriented at
an angle of 60°.
Hardwood trees have a different backscattering behavior because of the dif-
ferent size and distribution of their branches (with or without leaves) in the
canopy. The signature of a deciduous tree is very different from that of a conifer
tree, as can be seen for a walnut tree (Fig. 4.29), a very asymmetric tree. In this
case, a minimum of -13 dB was observed at a polarization angle of 60°, fol-
lowed by an increase to -9 dB at 85°. The decreasing part of the curve, between
0° and 60°, is due to the thick branches down to 45° from nadir. The second
part of the curve for polarization, above 60°, was related to the connection
angles between the trunk and branches. The modification of illumination
geometry indicated a different orientation and branch structure (Fig. 4.29).
Sieber (1985) concluded that the selection of an appropriate polarization with
an X-band and an L-band radar permitted the detection and identification of
deciduous and coniferous trees. The experimental data reported by Sieber
162 Microwave Imaging
decidous forests
oQ(dB)
-10
~ -12
0
-;;; -14
-=
.:2 -16
ta;::: -18
-=Q) -20
~ -22
-24
-26 height (m)
-28
0 1.25 2.50 3.75 5.00 6.25
a) horizontal polarization
decidous forests
oQ(dB)
-10
~
-12
0
-14
-;;;
-16
-=
.:2 -18
ta
-20
~ -22
~
-24
-26
height(m)
-28
0 1.25 2.50 3.75 5.00 6.25
b) vertical polarization
coniferous forests
oQ(dB)
-10
-12
~ -14
;:: -16
~
-=0 -18
·~-20
2-22
Q)
~-24
-26
height (m)
-28
0 1.25 2.50 3.75 5.00 6.25
C) horizontal polarization
Fig.4.26a-d. Penetration measurements at L-band radar for coniferous and deciduous forests.
a, c Horizontal HH polarization. b, d Vertical VV polarization. (Sieber 1985, with permission)
Applications 163
-10
-1 2
~
0
-14
x -1 6
c - 18
0
.~
-20
..,c'" -22
t::
ro -24
-26
height (m)
-28
0 1.25 2.50 3.75 5.00 6.25
d) vertical polarization
a)
tree
"
//1
radar antenna
height of radar
above gro und
. '"" f I
vertical heighr of
bcamltrce intersection
height of radar
above ground
\ 'I
IV Fig.4.27a-c. Experimental arrangement
I
for penetration measurements (Sieber
~\"""'
1985; courtesy of ESA Journal). a The posi-
tion of radar antenna and of the tree.
b Position of the radar central beam and of
the polarization direction versus the tree.
[C5S
c The INSCAT (Institute for Navigation's
'\
,,).=;)
Scatterometer) sensor mounted on a truck.
(1985) were collected in the Rhine Valley, Germany, with the INSCAT (Institute
for Navigation's Scatterometer) sensor mounted on a truck (Fig. 4.27c).
0° (dB)
-5.0
-6.0
-7.0
-8.0
-9.0
- 10.0
- 11 .0
-1 2.0
-13.0
- 14.0
(a) 0.0 25.0 50.0 75.0 100.0
0° (dB)
-5,0
-6,0
-7,0
-8,0
-9,0
- 10,0
-11 ,0
-12,0
- 13,0
po larization angle (degree)
- 14,0 I I I I I
(b) 0,0 25,0 50,0 75,0 100,0
c) d)
166 Microwave Imaging
of various species was performed in the frequency band between 4 and 8 GHz.
In this section, we discuss the imaging technique used for the detection of
knots in a spruce bolt of 17.5 cm diameter and 18 cm length. The actual dis-
tribution of knots is sketched in Fig. 4.30 where we observe three large knots
designated as a, b, and c. They are disposed in a radial position at about 900
between knots a and b and at 600 between knots band c. Knot b is externally
visible.
Figure 4.21 shows the experimental device for knot detection. It is composed
of three main parts: the microwave device with a wide band antenna, the con-
veyer for bolt rotation and translation, and the computer for equipment control
and for storage of experimental data. The image reconstruction is based on the
following dielectric parameters of the wood: the polarization ratio and
the phase difference of two components of the electrical field, one parallel to
the fibers and the other perpendicular to the fibers. The wave propagation phe-
nomena that take place in the sample are very complex and a simplified
hypothesis must be introduced to facilitate the understanding of these phe-
nomena. In a first approximation it is assumed that the received signal at the
antenna is the sum of two signals, the first emitted from the log and the second
emitted from the knots. Secondly, it is assumed that for the continuous case,
the sum can be replaced by an integral and the dielectric constants by a tensor
function.
Continuous wave radar measurements of the depolarization of reflected
waves were used to reconstruct the image. In the frequency range from 4 to
8 GHz, it was possible to obtain a discrete frequency spectrum. Using the
inverse Fourier transform, the discrete spectrum measurements were trans-
formed into a time delay spectrum. This spectrum shows the spatial distribu-
tion of the electrical field and of the polarization-related parameters that are
used for image reconstruction.
Applications 167
The images obtained with the polarization ratio (Fig. 4.31a) have less reso-
lution than those images obtained with phase difference (Fig. 4.31b), on which
the knots are better defined. An improvement of the image resolution was
obtained using iso-surface topographic slices calculated for the polarization
ratio and for the polarization angle between the two components of the elec-
tric field (Fig. 4.32a,b). Knots a and b are readily visible in both cases, while
knot c is readily visible only in Fig. 4.32b, where data based on the polariza-
tion angle were used.
Further research is needed for improvement of the image resolution and for
a better understanding of possible correlations between the dielectric tensor
Fig.4.31a-d. Iso-surface on a topographic slice. a Ipi The amplitude; b ythe argument (y, arg p).
c E The ellipticity angle which gives the phase difference between two wave components and
describes the degree of elliptical polarization. d r The angle between the two components of
the wave, corresponding to the tilt of the linear polarization. (Kaestner and Baath 2000, with
permission)
168 Microwave Imaging
a) real image
c)
Fig.4.32a-c. Microwave imaging of a knot in softwood lumber. a Real image of the board.
b Longitudinal-tangential microwave imaging of the section through AA. c Radial- tangential
microwave imaging of the section through BB. (Chazelas 1991, with permission)
and the physical parameters of the wood, such as the slope of grain and the
variation of moisture content. Today the technology is available to build a
microwave scanner that can operate at industrial rates. The only important step
to be fulfilled is the improvement of image resolution.
Applications 169
Fig.4.33a,b. Images of knots reconstructed with a the polarization ratio; b the phase differ-
ence between the two components of the electric field, parallel and perpendicular to the fibers.
(Kaestner and Baath 2000, with permission)
172 Microwave Imaging
dency. It is well known that the moisture in wood is composed of both water
and dissolved ions. For a hypothetical liquid containing both water and ions,
the variation of 10" versus frequency is different than that of pure water as can
be seen from Fig. 4.36. The bound water does not have the same behavior
as free water. The bound water reached a maximum value between 108 and
length length
Fig. 4.34. Microwave imaging of a softwood board (100 X 40 X 1000mm) (Choffel 1999, with
permission)
wavelenght (em .)
0
30 3 0(3
0
S'
0
.;;
:€
~o
P.-
/
<l)
;;.
/
'.g /
V.... /
S"/
/
10 100
frequency (GHz)
complex permittivity of water at 25°C
Fig. 4.35. Relationship between the permittivity and frequency for pure water at 25°C.
(Thompson 1989, with permission)
Applications 173
('
loss
t
factor
total
50 ionic conductivity
free water
bound water
/
I
........ -<.
1- - _-I'. -- ....
,I
Fig.4.36. e" as a function of frequency for a hypothetical dielectric material containing water
and dissolved ions. (Thompson 1989, with permission)
109 Hz, in the same domain in which free water reached a minimum. The influ-
ence of free water is more important than that of bound water at a frequency
higher than 1 GHz. This graph (Fig. 4.36) can help in selecting the appropriate
frequency for wood drying because of the capability of the electromagnetic
radiation, directed through the bulk material to quantify the water content in
real time (Thompson 1996).
From a large number of references we can see that the dielectric constants
were extensively used for measurements of the moisture content in wood and
wood-based-composites (Nanassy 1970, 1972; Norimoto and Yamada 1972;
Tiuri et al. 1980; Skaar 1988; Kuroda and Suzuki 1996; Yokoyama et al. 1999,
2000a,b).
Thompson (1989) found that the volume of water present in the dielectric
material can be deduced from the complex permittivity of the sample E*, and
can be expressed as follows:
(4.27)
where: Vm is the fraction volume of water; e,.* is the complex permittivity of
water, and Ed* is the complex permittivity of the dielectric material such as dry
wood substance. In this equation, all parameters are known except the volume
fraction of water, Vm'
For in-process quality control, King and Basuel (1993) developed a non-
contact microwave transmission technique for measurement of the moisture
content of fiberboard, based on the assumption that the variation of the atten-
174 Microwave Imaging
uation and phase of the microwave signal are linearly dependent on the dry
masses of solid components and of water.
The absolute moisture content, MC~w, deduced by the microwave technique,
is calculated as the ratio of the partial basis weights of the water present to the
partially dry wood weight.
(4.28)
where MC~w is the absolute fractional moisture content of the sample, mw is the
weight of water, and md is the weight of wood on the basis of the absolutely
dry fraction.
The total weight mtot.l can be deduced as the sum of the components:
(4.29)
The variation of the amplitude and of the phase as functions of md and mw
are given by:
(4.30)
(4.31)
where aj and az are experimental calibration constants for amplitude, and, a3
and a4 are experimental calibration constants for the phase.
After calculation, the expression of the moisture content as the ratio
between mw , the weight of water, and md, the weight of wood, on the basis of
an absolutely dry fraction, the moisture content is expressed by:
MC~w =mw/md = (a3LlA-ajLl<!»/(azLl<!>-a4LlA) (4.32)
The values of the coefficients a j az a3 and a4 are given in Table 4.11.
Figure 4.37 gives the variation of MC~w as a function of the gravimetric
moisture content measurements in the range of 0 to 20%. A breakpoint in the
variation of amplitude and phase were observed at 6% moisture content. Under
Table 4.11. Values of the coefficients al> a2, a3, a4 calculated by a least square regression for fiber-
board at constant temperature of 24°C. (King and Basue11993; courtesy of Forest Products J)
Moisture content md (g/cm2)' a 1 (dB cm2/g) a2 (dB cm2/g) a, (deg cm2/g) A4 (deg cm2/g)
'md (gtem2) is the weight of wood per unit area of board, on the basis of absolute dry fraction.
md = Pdt; where Pd (gtem') is the dry bulk density of the board of thickness t (em).
Applications 175
20
md=614kglm 3
16
~ me>6%+0.43%
e.... 12
~
~
], 8
a
()
4
me<6%+0.45%
0
0 4 8 12 16 20
me (by weighing) [%]
Fig. 4.37. Relationship between the moisture content determined with a microwave technique
and the routine moisture content measurement for a fiberboard of 614kg/m3 density, at 24°C.
(King and Basue! 1993; courtesy of Forest Products J)
6% moisture content, the water molecules are strongly bound to the chemical
constituents of wood and the attenuation of the microwaves is reduced and
exhibits a nonlinear behavior. Increase in the moisture content from 6% to the
fiber saturation point activates weaker bonds of hydroxyl groups of the hemi-
celluloses. At the fiber saturation point, free water is present in the cell lumina.
From the electromagnetic point of view, there appears to be little or no differ-
ence between the polarizabilities of weakly bonded and free water, so that I1A
and 11<1> are linearly dependent on the moisture content above 6%.
The influence of moisture content on the complex dielectric constants of the
fiberboard is shown in Fig. 4.38. The real dielectric constant £' is linearly
dependent on the moisture content, whereas the dielectric constant £" is non-
linearly dependent on the moisture content. During the processing, the
temperature of a fiberboard varies over a very large range (from 200 e after 0
emerging from the press, to 20 0 e in the final stage). The influence of temper-
ature on the amplitude and phase variations at different levels of moisture
content, ranging from 0.3 to 10.4%, is shown in Fig. 4.39. Both the amplitude
and phase variations increase linearly with the temperature. The phase
variations seem to be more sensitive to an increase of temperature when the
moisture content increases from 5 to 10%.
Microwave attenuation was used by Musial (1988) to study the degree of
flake orientation of OSB under the assumption that the physical characteris-
tics of flakes are similar to those of solid wood from which the flakes had been
produced. On the other hand, Musial (1988) found that the influence of resin
176 Microwave Imaging
2.4
2.0
0 4 8 12 16 20
1.0
0.8
0.6
0.4
0.2
4 8 12 16 20
mc(%)
and water repellent on the dielectric properties of the board can be neglected.
The third point is that there is no electrical interaction between the board and
the individual flakes. The dielectric properties of the flakes are not significantly
modified during board production. The anisotropy of the boards is expressed
by the ratio between the attenuation parallel and perpendicular to the direc-
tion of board formation. Considering the coordinate system x, y and z for the
board and x', y' and z' for the flakes as shown in Fig. 4.40a, the attenuation
anisotropy ratio is a linear function of flake orientation (Fig. 4.40b). This linear
relationship can be used in further development of a continuous nondestruc-
tive monitoring technique in the processing of boards.
The results previously described are based on the microwave transmission
technique. More recently, the development of monostatic transmission sensors
in reflection-aperiodic-open configuration (King 2000) permitted the con-
tinuous control of large dielectric boards. Volgyi (2000) proposed the utiliza-
tion of a multifrequency permittivity monitoring system, in which the
attenuation and the phase are measured and correlated with the moisture
Summary 177
1 ~I%
5 constant moisture content on amplitude
and phase variations, measured on a
fiberboard. (King and Basuel 1993;
courtesy of Forest Products J)
- 1,7%
1 0,3%
o !
30 40 50 60 70 80 90
mc=lU,4%
130
120
8,1%
110
~ 100
5,1%
~ 90
(3. 80
<1 70
60 tt~~~::======= 0,3%
1,7%
50
30 40 50 60 70 80 90
temperature [0 C]
content and density of the board and with the corresponding mechanical prop-
erties. Imaging of the distribution of the moisture content, resin proportion
and other parameters can be obtained for each board. Numerical analysis of
microwave measurements was based on the implementation of Kasa's circle fit
procedure (Corral and Lindquist 1998) that minimizes the errors of antenna
pattern measurements.
In the future, the holographic technique of a multifrequency microwave
method as generally described by Drobakhin (2000), for dielectric material,
could be an important advancement in the nondestructive control of large
wood-based composite boards.
4.5 Summary
o right
• left
o~--~----~----~--~~
o 0,2 0,4 0,6 0,8 rad
There are two basic microwave techniques, the transmission and the reflec-
tion technique. Signal analysis of probes is relatively simple and is related to
the measurement of amplitude, phase and the polarization of the waves. These
parameters are used for image reconstruction of wood structure. Today, the
areas that may benefit from microwave imaging techniques are related to the
internal defect detection of logs, lumber and wood-based composites, and, also
to the imaging of vegetation including leaves, stalks and stems under various
moisture and temperature conditions over a wide range of frequencies.
The imaging of wood structure can be obtained by scanning in the prox-
imity of the sample with a resolution given by the aperture size. The images
can be obtained in the far field or in the near field. In the far field, the small-
est detectable defect size is determined by the ratio d//.. >1, where d is the length
of the defect in the plane normal to the microwave vector and /.. is the wave-
length. The probes used to evaluate the properties of the medium can be either
scatterometers or radars, which measure the scattering properties, or reflec-
tometers, which measure the reflectivity of the medium due to its inhomo-
geneities and defects. The probes can operate in the proximity of the medium
as open-ended coaxial lines, cavity resonators or antennas. By moving the
position of the transmitter, an image of the medium can be obtained. While
microwave techniques have shown great potential for lumber inspection, their
applicability seems to be limited by the impossibility of identifying the nature
Summary 179
of the defects. To overcome this difficulty a vision system has been coupled to
microwave antennas.
The development of the microwave imaging technique is based on several
advantages such as:
- a noncontact operating system and relatively small size of antennas,
determined by the wavelength at microwave frequencies;
- fine resolution of measurements that can increase the ability of this
technique to detect defects;
- wave parameters, such as amplitude, phase and polarization can be
measured in real time.
The difficulties of applying microwave imaging to wood material arise from
inherent material properties, such as the anisotropy, heterogeneity, and the
presence of natural defects in wood. The main disadvantage of the microwave
system is that at a high rate of inspection the mechanical vibrations of logs or
lumber disturb the measurements of polarization. Technical limitations,
associated with the advantages of microwave imaging techniques, illustrate
the potential and capability of this method for new practical applications.
5 Ultrasonic Imaging
5.1 Introduction
IA
::'" ' ;" "
?- ..
..... ... ...
' . '.~" " "'-"',
.'
'.
, -,
..•.. "
:" ~;>::;'\
0,8 0,8
.'
.'
0,6 ." , '.
"
,,'
"
0.4
... - .. . ..,-
{" '~'
. ...
~
~
0.2
" ,
I .•
" >':,
. ~.
."'
0.8 . ~\
Fig.S.1a-c. Energy flux deviation shown on
" slowness curves, as a function of direction of
' -" .
0.6 ultrasonic wave propagation in oak, in three
anisotropic planes (Berndt, pers. comm.). a
Plane LR; b plane LT; c plane TR. The slowness
curves correspond to the quasi-longitudinal
waves (the internal curve) and to the quasi-
transverse curves (two external curves)
deduced from Christoffel's equation. The
o 0.2 0.4 0.6 0.8 1.2 L4 numerical data were selected from Bucur and
C) , 10""", (pslmm) Archer (1984)
perpendicular to the propagation direction. This is not the case for quasi-shear
waves, for which there is an important variation of color. This means that the
polarization vectors of quasi-shear waves are more sensitive to the interaction
with wood structural elements than is that of the quasi-longitudinal wave. The
examination of the interaction between wood structure and the propagation
phenomena has been discussed in detail by several investigators (Bucur 1980,
1995; Berndt and Johnson 1994; Berndt et al. 2000).
For media of interest to the geophysics community, the modeling of veloc-
ity propagation (Kline et al. 1994; Wang and Kline 1994) was performed using
either finite difference or finite element formulations. In wood science, pio-
neering works on wood structure imaging reconstruction by scanning, from
184 Ultrasonic Imaging
a) X Polari ation
along 8
(tran verse wave)
Polarisation
along r
(longitudinal, a c)
Polari ation
along IjI
(tran erse wave)
b)
Oak
,.' .
.::.'
( ... ."
....
0.5
......
0
-0.5 'r'
: ....
-1
"·1
Fig. S.2a-c. Three-dimensional
representation of acoustic
properties of oak (c). a Local
basis and color code. b
0.5 Variation of polarization angle
on slowness surface for oak.
0
(Bucur et al. 2001, with
C) -0.5 permission)
Basic Aspects 185
Fig. 5.3. Schematic representation of the Fourier diffraction theorem (Kak and Slaney 1988),
which relates the Fourier transform of a diffracted projection to the Fourier transform of the
object along a semicircular surface in the frequency domain. (Copyright 2001 IEEE)
after a few iterations. One can assume that the cross section of the object
consists of an array of unknowns (Fig. 5.3). After setting up the corre-
sponding algebraic equations for the unknowns in terms of the measured
projection data, the image can be reconstructed.
The algebraic algorithms used most frequently are represented by the
following acronyms:
- ART - algebraic reconstruction technique - on which each equation cor-
responds to a ray projection. The computed ray sums are a poor approx-
imation of the measured ones and the image suffers from significant
noise.
- SIRT - simultaneous iterative reconstructive technique - reduces the
noise of ART by relaxation and produces better images than ART. The
relaxation parameter becomes progressively smaller with increasing
number of iterations. A comparison of ART,SIRT,LSQ (least-squares) and
SVD (singular value decomposition) tomographic inversions for geo-
physical applications is given by McGaughey and Young (1990). ART and
SIRT are iterative inversion techniques, while least squares and singular
value decomposition are direct inversion techniques.
- SART - simultaneous algebraic reconstruction technique - combines
ART and SIRT and yields image reconstructions in one iteration, using a
Basic Aspects 187
P(t,B2)
Fig.SA Reconstruction of the image with fan beam projections collected in SI> S" etc. (Kak and
Slaney 1988); x,y coordinates related to the object in the space domain, ~1 and ~2 angles related
to the position of SI and S,; profiles PI (t, ~1) P2 (t, ~2)' etc. as a function of position and time t.
(Copyright 2001 IEEE)
- the number of sampling data along the receiver line, because the sampling
is limited and a nonzero sampling interval must be chosen. With the Nyquist
theorem, a low pass filtering operation can be introduced and the highest
measured frequency can be calculated from kmeasured = relT, where T is the
sampling interval.
- the limited views of the object under inspection, not estimated by the
Fourier transform, can degrade the image in certain directions.
35
2.50
25 2.00
......
~
..... 200 1.50
1.00
0.50
[mm]
Fig. 5.5. High resolution images of the transverse section of a tree (Platanus acerifolia)
(Martinis 2002, with permission)
location of the transmitter and the receiver. C-scan may be either of a reflec-
tion type or a through-transmission type. Modern digital computers can
extract the signal corresponding to each point of the work piece and produce
two dimensional images. Two-dimensional images in C-scan obtained with
modern digital computers require advanced signal processing. Volumetric
imaging can easily be obtained from a set of line scans covering an area of the
sample for the benefit of understanding propagation phenomena.
Equipment for Ultrasonic Imaging 191
Fig.5.6a,b. Earlywood and latewood in pine, imaged by ultrasonic reflection. a Ultrasonic and
photographic image superimposition at a 5-MHz area scan. b a at I-MHz area scan. (Berndt et
al. 1999, with permission)
b)
Fig.5.7a,b. Ultrasonic scanning of lumber by immersion in a water tank.. a Device for lumber
scanning. b Ultrasonic and real images of a lumber board (McDonald 1978; courtesy of Forest
Products Society)
collection and the authors noted that "scanning at relatively high industrial
speeds is feasible."
For direct transmission measurements, the direct coupling of the transducer
requires couplants that can be liquids (Birks 1972), grease, different adhesives,
air, etc. The air-coupled transducers were used in line production of particle-
board and plywood for the detection of blows and blisters with a speed of 90
m/min. The equipment for ultrasonic tomography available today is only for
laboratory measurements and consists of an ultrasonic generator, transducers
and an oscilloscope. The ultrasonic signal obtained from field measurements
require modern and complex processing (Berndt et al. 2000).
Equipment for Ultrasonic Imaging 193
a) b)
1 4 0 . . . - - - - - - - -- - ---, o
-10
120~---~------4
-20 .. .... - EV (dB)
1 00~~~·~~~~~~-4 -30 1-- - -EPV (dB) - - -
80 -40
60 i--- -
40~~~--_.~---~~
-50
-60
-70
~
);.I',~
·V- ~
.
-80
-90 .-
-100
\0 20 30 40 50 0 10 20 30 40 50
c) d)
350 , . - -- - - - - - - - - - - , o
300
---·----EV(dB) _
-+---------i--~--l ·20
--Energy - ---. EPV (dB)
250 ' -'-'- ' - Pulselenglh -t------:c---l -40
----- TOF·cenlro id :
200 -+---------:~--i--_l ·60
\ ...~
-v
ISO -j--- - - -- I I - -I/-- -I -80
~r ~
·100 I
I ~
·120
·140
o 10 20 30 40 50 o 10 20 30 40 50
Fig.5.8a-d. Parameters selected for ultrasonic scanning of board: time-of-flight centroid (TOF),
time-of-flight energy, time-of-flight amplitude, pulse length energy value (EV) and energy/pulse
(EPV). a Measurements of the TOF centroid through a yellow poplar sound knot. b Measurements
of the energy value and energy/pulse value through a yellow poplar sound knot. c Measurements
of energy, pulse length and TOF centroid through a decayed red oak board. d Measurements of
energy and energy /pulse value through a decayed red oak board.
Berndt et al. (1999) reported images obtained with relatively high frequency
ultrasonic transducers such as: 1 MHz focused transducers and 5 MHz unfo-
cused transducers using the device depicted in Fig. 5.9a. The line scan in the
radial direction of the sample is shown in Fig. 5.9b, in B-scan technique. The
194 Ultrasonic Imaging
e) f)
450 - . - - - - - - - - - - - , o -r-~-~-..-----,.-...,
..... ... EV (dB)
400 +-~- --Energy ·20 .
--EPV IdB)
350 ---- .. -- Pulse Length
300 --- - - TOF -cenlloid
·40 -1- - - - - - - - -1
Fig. 5.8e,f. e Measurements of energy, pulse length and time-of-flight centroid through a crack
in red oak lumber. f Measurements of energy and energy Ipulse value through a crack in red oak
lumber. (Kabir et al. 2001, with permission)
"
.. '.
-~
10
c) radial posi1ion (rnm)
Fig.5.9a-c. Equipment for contact scanning with immersion technique. a Ultrasonic scanning
setup. b Scan line in radial direction at 5 MHz. c Total reflected energy and peak reflection arrival
time vs. scan position. (Berndt et al. 1999, with permission)
196 Ultrasonic Imaging
a) b)
Q
c)
~,;,
-
e e
~
\.
!- .•• ,
»1' • • • ')
~".II.,
_I.....I-i'''''\!'_...............-
~. ~ ...... "." . . '\.
TableS.I. Ultrasonic energy transfer in various materials in non contact mode. (Bhardwaj 2000; with
permission)
' The impedance is expressed in Mrayl (1 RayJ = 1 kg m-' S- I). The impedance of air is 440 MrayJ and the im-
pedance of water is 1.5 Mrayl.
198 Ultrasonic Imaging
" .....
a) b)
Fig.5.11a,b. Air-coupled ultrasonic waves in a wooden plate. a Two spherically focused piezo-
electric transducers coaxially aligned. b The focused transducer is coaxially aligned with a fiat
piezoelectric transducer. (Fortunko et al. 1991; copyright 2001 IEEE)
with a thickness between 3 and 25mm. The first configuration using two coax-
ially aligned, spherically focused transducers, having the axes of symmetry
parallel, is shown in Fig. 5.11a. The central plane of the sample contains the
foci of both transducers. The second configuration (Fig. 5.11 b) used only one
focused piezoelectric transducer and one unfocused transducer. The focal
point of the spherically focused transducer was located at the surface of the
wooden plate in front of the unfocused transducer. In both cases, the angle of
incidence (9) is variable to allow the measurement of the time of flight and of
the amplitude of the ultrasonic signal as a function of this angle. It is impor-
tant to note that transmitting transducers must be driven by a high-power,
unipolar pulse generator (450 V output voltage, 3Q impedance and 2000 V
peak-to-peak voltage applied to the transducers). A low-noise preamplifier and
four precision attenuators (total range 132 dB) were used to minimize the noise
factor.
The design of higher frequency noncontact probes (>500kHz) is difficult
because of the short wavelength and high absorption in air, and only two solu-
tions were considered efficient: piezoelectric ceramics elements modified to
optimize its transduction in air and capacitive transducers.
Recent advances in high frequency piezoelectric transducers (Bhardwaj
et al. 2000), in the frequency range of 200 kHz to 5 MHz, coupled with a
high dynamic range-nanosecond accuracy ultrasonic analyzer that is able to
provide 100-dB extra gain, 30 dB of which is used to compensate the non con-
tact mismatch and 70dB to overcome air-material acoustic impedance mis-
Equipment for Ultrasonic Imaging 199
....--.--.. •
•,
~
•.
•
"'''' 1 ' '"
Ii
Ii
Fig. 5.12. Different ultrasonic scanning modes of sheet products. (Buckley 2000, with permission)
5.4 Applications
Topics covered in this section include imaging at macroscopic level of the inter-
nal structure of standing trees, of lumber, and of defects in wood-based com-
posites. Acoustic microscopy imaging is not discussed in this book.
In considering the application of ultrasonic imaging, it is important to note
that for wood the transmission technique seems to be the most appropriate.
Since resolution in an ultrasonic image is dependent on frequency, it is impor-
tant to note that the size of the sample to be inspected, the type of the trans-
ducer and the band width must be chosen with care, keeping in mind the
propagation phenomena and the anisotropy of the material.
Most ultrasonic inspection and imaging are done at frequencies ranging
from 50 kHz to 1 MHz. As the frequency increases above 1MHz, the image res-
olution increases but the signal attenuation increases also. Ultrasonic images
are degraded by blurring due to practical limitations imposed by beam diam-
eter and pixel size. Digital processing methods may be used to improve the res-
olution. The spatial resolution, at which the original data were acquired, limits
the improved resolution in the processed image. The resolution inherent to
acoustic images is basically determined by the beam diameter and by the pixel
size. Nyquist's theorem explains the effect of the pixel size on the image reso-
lution. The pixel size is inversely proportional to the resolution. In order to
support the spatial resolution, the beam must be spatially sampled at less than
half the dimension of the pixel size or, in other words, the pixel size of the
image must be one half of the ultrasonic beam diameter.
Applications 201
_--4
RF amplifier ....
P Texas
Instruments
Fig. 5.13. Inspection of a veneer sheet with Lamb waves. Above Experimental setup for Lamb
waves launched by Herzian contact in a multilayered structure composed of a veneer sheet glued
on a particle board. Below Transducer for Lamb waves measurements. (Grimberg et al. 2000, with
permission)
202 Ultrasonic Imaging
where T is the travel time of the ultrasonic wave from the source (S) to
the receiver (R ), I is the ray path length, v is the velocity and p = l/v is the
slowness.
The mathematical background of the inversion algorithm is treated in
reference books and articles and will not be discussed in this book because
of space limitations. As noted by Socco et al. (2001), the procedure can be
described as follows: "the velocity calculation was performed with SIRT,
starting from an initial arbitrary velocity distribution model, computing the
forward model travel times, comparing the calculated travel times with the
measured time, and correcting the initial velocity model. These steps are
repeated until the fitting between calculated and measured travel times is
considered good enough:'
The flow chart for data acquisition, analysis, and interpretation is shown in
Fig. 5.14. After the processing of ultrasonic data, the tomographic image was
compared with the real (photographic) image of the transverse section of the
tree. For this purpose, a disk of 40 cm thickness was cut from the tested tree.
From this disk, cubic specimens were cut, on which ultrasonic velocities were
measured and used to construct an alternative of velocities distribution. This
Applications 203
Trees
cutdown
Good
signal/noise
ratio
Fig.5.14. Flowchart for the tomographic imaging of the transverse section of a tree. (Socco et al.
2001, with permission)
(5.2)
204 Ultrasonic Imaging
25 ~-----------------------------------------------------,
20 +--------------------------------------------
~
ii
~ 15+------------------------------------
!
10+-------------------------------- - - - -
o
2.00 2.26 2.53 2.79 3.06 3.32 3.59 3.85 4.11 4.38 4.64 4.91 5.17 5.44 5.70
Pre nel r.y classes ( em )
Fig.S.IS. Fresnel ray distribution calculated at S4kHz on the hackberry trunk (Celtis australis)
shown in Fig. S.14. ( Socco et al. 2000, with permission)
The Fresnel ray distribution is given in Fig. 5.15 for measurements per-
formed on a 63-year-old hackberry tree (Celtis australis) with very complex
trunk shape caused by previous wounding and pathological attacks (Fig. 5.16).
The most probable Fresnel ray class is at 5 cm, which corresponds to the best
image resolution. A good agreement was observed between the three images,
that is, the photographic image (Fig. 5.16), the tomographic image of the trans-
verse trunk section obtained from 120 independent velocity measurements
(Fig. 5.17), and the reconstructed image with measured velocities on cubic
samples (Fig. 5.18a). It should be noted that the sharp velocity contrasts are
smoothed by the tomographic inversion.
For the illustration of the simultaneous variation of the velocities in the L,
Rand T directions, the specimens presented in Fig. 5.18b were used to measure
the corresponding velocities. The variation of velocity values versus the posi-
tion of the specimen is illustrated in Fig. 5.19. Comparing these results and the
values of the velocities employed for the tomographic imaging, it can be con-
cluded that the tomographic images match the images constructed from veloc-
ity measurements on the individual blocks.
Another example is shown in Fig. 5.20, in which the tomographic image of
the transverse section of a standing tree (Platanus acerifolia) of 450mm diam-
eter with an important decay zone is presented. The velocities measured in the
RT plane range between 1000 and 1800m/s. In the central, decayed zone, the
velocities are relatively low compared to sound wood. This is probably due to
the low density of the tissue affected by the fungal attack. The low velocity
peripheral zone observed on the tomographic image is caused by the experi-
mental configuration and is likely to be an artifact of the inversion algorithm,
because this zone lies outside all measured ultrasound paths.
Applications 205
Fig. 5.16. The complex shape of the transverse section of the trunk of hackberry (Celtis australis)
(Comino et al. 2000, with permission)
600 2.00
1.80
500
1.60
400 1.40
E 1.20
§. 300
1.00
200 0.80
0.60
100 D.40
0.20
100 200 300 400 500 600
[mm]
Fig. 5.17. Tomographic images of the transverse section of the trunk in Fig. 5.16 obtained with
120 independent velocity measurements. (Comino et al. 2000, with permission)
206 Ultrasonic Imaging
a)
km/s
600
1.80
500
1.60
400 1.40
E 1.20
.§. 300
1.00
0.80
0.60
0.40
100 200 300 400 500 600
b) [mm]
Fig. 5. 18a,b. Reconstructed image with velocity values measured on cubic specimens and posi-
tions of the specimens in the trunk section. Above Reconstructed tomographic image. Below
Cubic specimens. (Comino et al. 2000, with permission)
Applications 207
2
1.8
1.6
'OJ'
~ 1.4
Co
1.2
~
Q
l 0.8
0.6
0.4
~ Z~ ~ ~ ~ ~ ~ ~ ~ ~
sample labels
si Uj
z z
fLI
Fig.5.19. Measured ultrasonic velocities and calculated velocities for two orthotropic directions
Rand T on selected specimens shown in Fig. S.lS. (Socco et al. 2000, with permission)
1.80
1.60
1.40
1.20
1.00
0.80
0.60 Fig.5.20a,b. Ultrasonic
0.40 tomographic reconstruction of
the transverse section of a
standing tree (Platanus
acerifolia) 40cm above ground
level. (Martinis 2002, with
b) ImmJ permission)
208 Ultrasonic Imaging
e22
~
a) e33
-
- VI.--' r-::;
--
<
V :/
..... ..-' 11\
r-
~1 l\'
-
-_
1/
-t- ........ r, r-; k '- 1--
-f- ...... =' f- ......
I-
--
-.....: ::--;:
Fig.5.21a,b. Reconstruction of the image of boards with loose and tight knots in a pine sample,
based on the measurements of stiffness CLL = Cll CRR = C22and Cn = C33 and related angles on
cubic specimens compared with the real image of boards. (Chazelas et al.1988, with permission)
veneer sheets of I-mm thickness were glued onto the two faces. The delami-
nations were simulated by inserting rayon foils of 50 J.lm X 0.5 mm as shown in
Fig. 5.23a. The density of the wood composite was 780kg/m\ the Young's
modulus in the longitudinal direction was 1.4 x 10 1O N/m 2, and the corre-
sponding Poisson ratio was 0.14. The velocity in sound zones was 4300±90m/s.
In a zone with delamination, the velocity measured in the veneer layer was
significantly lower, ranging from 1100 to 2128 m/ s. The map of the delamina-
tions is shown in Fig. 5.23b. The location of each delaminated zone is clearly
visible.
210 Ultrasonic Imaging
a) I 20mm
200mV l100mV
SOOmV l100mV
Fig.5.22a,b. Ultrasonic images obtained in C-scan mode by immersion and in reflection mode
of the structure of a pine specimen. a Longitudinal-radial section with knots, the pattern of the
earlywood and latewood and the slope of grain deviation around the knots. b Transverse section
with typical pattern of the annual rings and the compression wood. (Neuenschwander et a1.1997,
with permission)
a) 6cm
'n
<:>
5 10 10 10 4 (em)
b)
•••
:."
:-"
:.•
:.•
.•
'.
:.'a
.
:•
.,.'
."
""
"
:•
"
,.."
.
"
.'0-
"
•
:
"
"
"
."•"
Fig. 5.23a,b. Imaging of delamination in a composite with Lamb waves, The composite was
produced with two beech veneer sheets glued to the two faces of a particleboard layer. a Position
of the delaminated zones. b Image of the delaminations obtained with Lamb waves. (Grimberg
et al. 2000, with permission)
212 Ultrasonic Imaging
20mm
1000 mV 5100 mV
Fig. 5.24. Imaging of holes in medium density fiberboard of 18-mm thickness in C scan mode,
by the immersion technique with O. S-MHz transducers. The image was superimposed on
the geometry of the specimen. The position of the second hole from the right is well defined
(Neuenschwander et al. 1997, with permission)
particleboard, where the elastic constants in the direction of the board for-
mation and in the corresponding perpendicular direction are similar, and only
the elastic constant thickness direction of the board is relatively low (Bucur
1992).
The ultrasonic imaging of voids in medium density fiberboards of 18-mm
thickness is shown in Fig. S.24. The voids were simulated by holes of different
diameters bored in the fiberboard. The image was obtained in C-scan mode,
by a transmission technique, with the specimen immersed in water and with
broadband transducers of O.S-MHz central frequency. In the future, for on-line
quality assessment, research on the development of dry or noncontact scan-
ning techniques in a large frequency range must be performed.
ultrasonic inspection is of great interest for electrical companies all over the
world because of its handling ease and safety. The ultrasonic propagation char-
acteristics measured on poles (Arita and Kuratanik 1984) and especially the
tomographic imaging (Tomikawa et al. 1985), readily interpreted by the field
operators of electrical companies, are excellent alternatives to the old percus-
sive hammer method.
The attempt to reconstruct ultrasonic tomographic images is based on the
assumption that the ultrasonic waves propagate in a straight line and that the
sound speed greatly decreases in zones where decay is present. The fan-beam
capture of ultrasonic time-of-flight was used and 324 points were employed to
reconstruct the image using 78-kHz transducers, for which the wavelength was
2 cm. This experimental arrangement was completed with 97 projection data
of parallel beams, obtained by interpolation of the fan-beam arrangement. This
large number of data was deconvoluted using the technique of filtered back-
projection. The computational time for producing an image was about 1 h,
which is too long for practical use. To reduce the time requirements, the recon-
struction of the images must be done in hardware. It is necessary to develop
an on-line system that enables rapid reconstruction for in-situ pole inspection.
5.5 Summary
The main purpose of this chapter has been to identify the breadth of applica-
tions of ultrasonic tomographic reconstruction techniques of wood at macro-
scopic levels. Ultrasonic tomography is a diffraction type tomography that is
noninvasive and is safe at low energy levels. Different types of ultrasonic waves
can be used for wood imaging, but the most common are bulk longitudinal
waves. Ultrasonic images can be reconstructed from all characteristic para-
meters of the wave: time of flight, amplitude, frequency spectra of the wave-
form, the phase, etc. The energy distribution and energy flow are important
parameters for enhancing the image contrast. In this chapter, we have dis-
cussed very recent results related to tomographic imaging with the time of
flight of ultrasonic waves.
There are three main types of algorithms that can be used to form tomo-
graphic images from ultrasonic data: transform techniques, iterative tech-
niques and direct inversion techniques. For the algebraic reconstruction
algorithms (ART), each equation corresponds to a ray projection. The sums of
the computed rays are a poor approximation of the measured ones, and the
image suffers from significant noise. The SIRT (simultaneous iterative recon-
struction technique) reduces the noise of ART and produces better images. The
factors that limit the accuracy of the images obtained with diffraction tomo-
graphic reconstructions are related, firstly, to the theoretical approach to the
approximations in the derivation of the reconstruction process and, secondly,
to the experimental limitations. High-resolution images have been obtained
214 Ultrasonic Imaging
for small, clear specimens of different species, for standing trees, for lumber
and for wood-based-composites. Two-dimensional imaging requires advanced
signal processing and modern digital computers.
The equipment for ultrasonic imaging can operate in mode A, Band C in
contact and non contact (or air-coupled) scanning, using frequencies ranging
from 50 kHz to 5 MHz. As the frequency increases above 1 MHz, the image res-
olution increases, as well as the signal attenuation. The resolution inherent in
acoustic images is basically determined by the beam diameter and by the pixel
size.
The piezoelectric transducers used for ultrasonic tomographic imaging can
operate in direct transmission mode, inducing bulk or surface waves. Conven-
tional piezoelectric transducers in the range of 50 kHz to 1 MHz are commonly
used for data acquisition. Air-coupled ultrasonic transducers are mainly used
for inspection of wood based composites like veneer or low thickness fiber-
board and particleboard. Ultrasonic tomography has a very large field of appli-
cations in decay diagnosis of standing trees, in parks and public gardens in
the imaging of lumber structure to reveal knots, grain deviation, cracks and
compression wood and in imaging of defects in wood composites, such as
delaminations between layers and voids.
6 Nuclear Magnetic Resonance
6.1 Introduction
Nuclear magnetic resonance imaging was first developed for medical applica-
tions. However, since 1980, an increasing number of researchers have reported
applications of this technique to materials science using standard NMR spec-
trometers, in which modifications had been made, and computing facilities had
been adapted to produce images. Every measurement in conventional NMR
spectroscopy, such as chemical shift, nuclear spin relaxation times, dipolar cou-
plings and the spectrum of translational motions can, in principle, be trans-
lated into a spatial resolved contrast and used to produce images. The major
impediment to the extension of this technique was and still is the price of the
equipment.
Nuclear magnetic resonance imaging offers to wood science and wood
processing a nondestructive, noninvasive and noncontact technique to study
wood structure characteristics, such as growth rings, the sapwood/heartwood
boundary, rays, reaction wood, resin canals, knots (Bucur 1990), wounds,
wetwood, decay by fungi, etc. Images obtained from magnetic resonance pro-
perties of iH, \3e or other nuclei provide a tool for in vivo analysis of trees.
This technique is useful in evaluating normal events in the life of trees as well
as different perturbations induced by external factors, such as climate, diseases,
etc. Tree vitality as indicated by the sapwood/heartwood ratio can also be
studied. Water distribution within the growth ring, in roots, branches and
leaves, diurnal or annual periodicity of the water content in the stem can be
determined in vivo. Wood drying is an important field of application of mag-
netic resonance imaging. Another research aspect is related to the distribution
of preservatives in wood or to the decomposition of wood by fungi and/or
other pathologic agents able to induce a modification of the moisture content
of wood (Birkeland 1990; Kucera 1990). When combined with localized spec-
troscopy, nuclear magnetic resonance imaging provides a unique experimen-
tal tool for the study of the structure of wood.
Knowledge of the fundamentals of nuclear magnetic resonance technique
is critical to the productive use of this technology for wood, which is a porous
material. The physics of magnetic resonance phenomena is well understood
and is described in various reference books. Amongst these books and review
articles, one can mention the books written by Abragam (1961), Talpe (1971),
Slichter (1978), Gadian (1982), Mansfield and Morris (1982), Schau (1984),
216 Nuclear Magnetic Resonance
Callaghan (1991), Gil and Neto (1999), and Lambert et al. (2000). The princi-
pal phenomena are identified in the literature by their initial letters such as:
NMR for nuclear magnetic resonance and MRI for magnetic resonance
imaging. This chapter reviews methods and applications of the NMR technique
to wood considered here as an anisotropic porous material, with emphasis on
imaging and spectroscopy techniques developed specifically for wood.
Table6.1. NMR properties of some nuclei (Gadian 1982) (© DG Gadian 1982, reprint from
Nuclear Magnetic Resonance and its applications to living system by David G Gadian (1982) by
permission of Oxford University Press)
the applied magnetic field. In the absence of a magnetic field, the average
sample magnetic moment produced by the nuclear magnetic moments is zero
because the spins are randomly oriented.
Quantum mechanically, a spin 1/2 nuclei can only align parallel or anti-
parallel to the field. Boltzmann statistics gives the fraction of nuclei in each
state with more nuclei in the lower energy state, with the nuclei magnetic
moment parallel to the field, and anti-parallel to the field direction There will
therefore be a net sample magnetic moment in the direction of the field.
Because of their angular momentum, the nuclei are subjected to a torque and
they precess like a gyroscope around the direction of the field. The frequency
of precession is the Larmor frequency 0)0 and is proportional to the gyromag-
netic ratio of the nucleus "I and field Bo as defined by:
(6.1)
The frequencies associated with conventional NMR techniques are usually
in the range of 100 to 600 MHz. NMR frequencies can be identified with the
absorption of microwave photons of energy h 0) (where h is the Plank's con-
stant = 6.6255 x 10-27 erg. s = 6.6255 x 10-34 N ms), causing transitions between
the energy levels associated with the quantum mechanical states of the nucleus
in a magnetic field as:
~E=h "(Bo (6.2)
where ~E is the difference in energy between the two states.
The nuclei resonance frequency is therefore directly proportional to the
local magnetic field experienced by the nuclei. However, the applied Bo field
also induces electronic currents in the surrounding atoms, producing a further
small shielding or screening field O"Bo. The effective field Beff at the nucleus is
therefore given by:
Beff =B o (1-O") (6.3)
0" is known as the shielding or screening constant and depends on the
atomic environment of the individual nuclei in a chemical compound.
The NMR frequency for an individual nuclei is therefore
0)0 = "I Bo (1- 0") (6.4)
where 0" has typical values from 10-6 to 10-3 •
Different nuclei in a chemical structure will therefore have slightly
different NMR frequencies, depending on their chemical environment.
Resonant frequency is defined by the chemical shift 0 (in ppm = part per
million) and is expressed as a fractional change in the frequency measured.
For example, the NMR spectrum of acetic acid CH 3COOH gives rise to two dis-
tinct frequencies for the protons in the CH3 and COOH parts of the molecule.
The zero value of the chemical shift for IH and 13C spectra is defined by
comparison with measurements on tetramethylsilane. The chemical shift is
independent of the magnetic field strength. The application of a small radio
218 Nuclear Magnetic Resonance
whole system will rotate and will have the new coordinates X', Y' and Z'. The
rotating magnetic field is produced by a radio frequency (RF) oscillator. No
effect is seen on Mxy, but a torque is exerted on Mz that will thus precess about
the X' -axis and about the applied field Bo direction. Rotation of Mz will induce
a voltage in a detector coil situated along the Y-axis, and resonance will be
detected.
The maximum induction signal, induced in the detector coil, will occur
when the pulse of RF energy at the Larmor frequency is sufficiently long to
rotate Mz at 90 into the XY plane. On removal of the Bl field, the nuclei will
0
precess with their spin perpendicular to the applied static field Bo. However,
the nuclei will experience small local fields that alter Bo. These local fields are
both positive and negative, so that some spins precess faster and some slower
and are no longer in phase. Their magnitude reduces to zero as exp( -t/T z ),
where Tz is the transverse relaxation time.
The thermally induced fluctuation in the local field associated with the
motion of atoms also produces components of the local field at the Larmor fre-
quency. This causes net magnetization. Mz returns to its equilibrium value as
Mz = Mo[1 - e(-tlT1)j, where Tl is referred to as the longitudinal relaxation or the
spin-lattice relaxation time.
The characteristic relaxation times for a decay of the Z and XY components
of the magnetization T I and Tz are strongly dependent on the environment of
the excited spins. The effects of relaxation and precession are combined in the
Bloch equations describing NMR (Abragam 1961). In general, Tl differs from
Tz because Tl involves a transfer of energy from the spin system to the sur-
roundings, while Tz describes the energy concerning loss in phase coherence
within of the spin system.
Fig.6.1a-d. Macroscopic description of magnetic moment of the nuclear spin system. a Nuclei
are in equilibrium with the static magnetic field Bo. Mo is the overall magnetic field vector; Mz
is the static magnetic field along the axis Z; Mxy is the rotating magnetic field in the XY plane;
e is the angle of precession. b Rotation of Mz in the presence the of Bj magnetic field generated
by the Larmor frequency. Mz is tipped away from Z direction by B field. c Loss of phase coher-
j
ence of spins after the extinction of the Bj field, leading to a reduction in precessing Mxy as
MXY(o)e-tiT '. d In practice, decay is observed because nuclei experience different fields in different
molecules and shift the effect that can be measured. The local variation in the field Bj is observed
due to the spin-spin interaction. T, is generated by sample properties (pore size, chemical
bonding of H,molecules)
a) b) Z'
Z
I
, ,.Mo
Bo
static field
Bo 1 Mx
r
o I
~ y I )h/ X,
X X'
d) t:C
c) 1 ~
Bo r;"
90° ~
!4.
'"
j '"o
...,
'" T"i FID Z
s::
:;.::I
time
8
X' i·
OQ
spins components ABC FID=free induction decay
tv
....
\0
220 Nuclear Magnetic Resonance
I
I T2
I
I
I Tc=lIwo
10- 5 L,-_ _ _ _ _ _-lII..-_ _ _t_im_e_(S_ec_On_d_es-,-l
10- 12 10-9
Tc (s)
Fig.6.3. Relaxation time as a function of molecular correlation time Teat typical frequencies
associated with spin motion. (Pfeffer and Gerasimowicz 1989, CRC Press, with permission)
scale, as shown in Fig. 6.3 for these relaxation processes, can be quite different.
Slow motions in the kHz range cause the spin to lose the precessional phase
coherence determining T2, while protons are associated with fluctuations of the
magnetic field, at the Larmor frequency (typically many hundreds of MHz) are
required to restore the spin population to equilibrium along the Bo direction
determining Tl'
Polymers of molecular weight between 10,000 to 25,000Da may have relax-
ation times on either side of the minimum of the curve, depending on
the Larmor experimental frequency. For molecules having relaxation times
Tc >10-9 s the relaxation times T1 and T2 are different. In contrast to the
relaxation of proton spin magnetization in wood samples, on which one finds a
multi-exponential decay time constants ranging between less than 1 ms to
several tens of ms, relaxation in bulk water is a single exponential process: T1 =
T2 = 2.4 X 10-9 m 2/s at 25°C. Diffusion measurements are of particular interest
in wood science and industry. For this purpose, we have to study the effects of
field gradients and of static inhomogeneities in field on spin-spin fluctuations.
By using field gradients perpendicular to the X, Y or Z axes, one can make
measurements on slices in all directions, using familiar computed tomographic
techniques. Three-dimensional information can be obtained from all points of
the space available for object imaging.
The field gradient is also important in determining the diffusion time
because during the measurements the water molecules diffuse in and out of
the measured section. Only molecules remaining with the section contribute
222 Nuclear Magnetic Resonance
to the magnitude of the pulse echo in a 90°-180° sequence. This enables the
diffusion coefficient in the direction of the field gradient to be measured. In
practice, for imaging improvement, Chang et al. (1999) demonstrated the utility
of the sophisticated sequence of switching of pulse gradients in all directions.
Interestingly, any reduction in precessing the Mxy component due to field
inhomogeneities can be reversed by applying a 180° pulse as indicated in Fig.
6.4. As the spins represented by the components ABC continue to precess rel-
ative to each echo by the same amount, they all come together at a time 2 t, to
give what is called a spin echo signal. However, alternatively, the effects of a
static inhomogeneity in the field can be reversed. Random processes such as
the spin-spin fluctuation affecting T2 and any diffusion of spins from one
region to another in an inhomogeneous field cannot be reversed. In a homo-
geneous field, this technique enables T2 to be measured, while in a field gradi-
ent the decay enables the diffusion content to be measured along the field.
For spins diffusing in a field gradient, the pulse echo sequence is determined
by the spin diffusion length (2 D t)1I2. If the length is known from the struc-
tural features, we can measure T2* to determine the diffusion coefficient D, or,
if D is known, the diffusion length, which can be related to the dimensions of
cell structure.
The NMR technique can also be used to investigate the diffusion of water
both along and perpendicular to the fibers. This is also achieved by NMR mea-
surements in a field gradient. In practice, during the sequence, the water mole-
cules can diffuse out of the structural region. If an interaction takes place
between the water and the cell wall, the alignment of spins can be destroyed.
This limits the time the spin can remain in a coherent state (the T2 time) so
that the diffusion time can be deduced and the pore size can be determined.
Moreover, a long diffusion time can be related to the diffusion measurement
along the fibers and can be used for the calculation of porosity.
Brownstein (l980), using NMR data published by Hsi et al. (l977) for
wood specimens at 38% moisture content of Thuja occidentalis, explained the
observed nuclear relaxation rates in terms of a strong relaxation. The deduced
relaxation times are shown in Fig. 6.5. For these specimens, the relaxation time
T2has two components, a slow relaxation time T2s of the order of a few tens of
ms and a fast component T2f of the order of a few ms. T2f is identified to be the
same as T\. Similar data were reported by Byrne et al. (l986) for Pinus radiata,
with two values ofT2, 0.8 and 10ms measured at 60 MHz. When, the specimens
were saturated with water, T2s increased considerably, while T2f leveled off at
about three times its initial value. To explain the phenomenon, Brownstein
(1980) proposed a model entailing (we suppose) diffusion of free water present
in the lumina and a monomolecular layer of water near the cell wall (corre-
sponding to a 6% wood moisture content), which is illustrated schematically
in Fig. 6.6.
The model consists of two water layers of thickness a and b with diffusivi-
ties D] and Du, where D] « Du. The strong magnetic relaxation is supposed to
be present only in region I. Calculations demonstrated that in region II the
Basic Aspects of NMR Imaging 223
90° pulse
--~~------------~~~~~--------~r-~ Bt - O
____~-4________~~____________~--~------------------~----~r_----~~r;n~
21
linle
Fig.6.4a-c. Relaxation time as a function of pulse sequence time. a Reduction of precessing Mxy
component due to field inhomogeneity can be reversed by applying a 1800 pulse. b The spins rep-
resented by the components ABC continue to precess, and they come together at a time 2 t, cor-
responding to the spin-echo signal. c For spins diffusing in a field gradient, the pulse-echo
sequence is determined by the spin diffusion length. If the diffusion length is known from geo-
metric structural features, one can determine the diffusion coefficient from T,* measurements
Fig. 6.5. Theoretical and experimental data
80' o
of TIf and T" as a function of moisture
o content. (Brownstein 1980, Academic Press,
70
T2s(ms) o with permission)
60 o
50
o
40
o
30
o
. data
20 • p
o theory
10
OL-------------------------
t3
T1f(ms) 000000
i:,.O"d 0 c
2
.....'19-
00 0.2 0.4 0.6 0.8 1.0 1.2 1.4
gram Hplgram wood
b)
water in lumen cellular material
--+II--~~_-_-_-_-+I=a=====.I-;-b-- X
Fig.6.6a,b. Physical model proposed by Brownstein (1980) for a wood cell, to explain the diffu-
sivity of water in wood, with 1-D geometry of the density of proton magnetic moment. a Free water
and bound water in a cell. b Diffusion model. Region I is defined by the thickness x = a and diffu-
sivity D] and corresponds to free water in the wood, which corresponds to 0.38 g H,O/g wood. When
only a monomolecular water layer is present, we have D]« Dn; a2/D] = 25.8 ms and a $0.26 /lm. The
diffusion process is inhibited within this layer. Region II is defined by the thickness x = b and the
diffusivity Dn and corresponds to the bound water in the cell wall. The proton distribution is con-
stant and Dn» D] and Dn = 2.4 x lO-scm'/s. (Academic Press, with permission)
Basic Aspects of NMR Imaging 225
Table 6.2. Diffusivity (D) in some species determined with the NMR technique, compared with
classic techniques
proton density distribution was a constant for each mode under consideration
and that the relaxation time was dependent on a2/DI> corresponding to the dif-
fusion term for water molecules across the thickness a. This parameter deter-
mines the time scale and "does not affect the form in which the relaxation time
varies with the ratio b/a:' The model requires that the ratio of the two diffu-
sivities DdDn = 1/900. If Dn is taken to be 2.4 X 1O-9 m2/s, for water, the para-
meter a2/DJ has the value of 26ms and consequently a < 0.26~m. This means
that region I consists of rather a thin layer of water molecules.
Like all physical parameters of wood, the diffusion of water is anisotropic.
MacGregor et al. (1983) calculated the diffusion coefficient of water molecules
under equilibrium conditions (static or self-diffusion) as a function of the
angle e between the longitudinal anisotropic axis corresponding to the cell
axis and the field gradient (Fig. 6.7).
The coefficients parallel and perpendicular to the cell axis are given in Table
6.2. The ratio between the two coefficients reflects the anisotropy of wood.
Water diffusion in sugar maple is more anisotropic than in Douglas fir. In the
direction parallel to the cell axis, the diffusion coefficient is smaller than that
of bulk water, probably because the water in wood (sap) contains different
magnetic ions (Mn, Fe, eu) and is different from bulk water and also because
the cell walls restrict diffusion.
Lindgren (1994) summarized in a very concise manner, the principles ofthe
NMRmethod:
1. the random orientation of the magnetic moment of hydrogen nuclei in the
absence of a field;
2. the lining up of nuclei when a magnetic field is applied;
3. the 90° tips on magnetization on applying a radiowave field;
4. the removal of the radiofrequency signal, and subsequent generation of a
radiowave by the precessing tipped magnetization frequency and amplitude.
Several methods for inversion recovery to measure Tl and T2 are used,
including saturation recovery and progressive saturation. The following
describes the inversion recovery pulse sequence as given by Hailey et al. (1985).
It was noted previously that the magnetic moment precesses at a frequency
proportional to the local magnetic field strength. The tipping at 90° and at 180°
depends on the strength and duration of the pulsed radio frequency signal. A
226 Nuclear Magnetic Resonance
l.5rr---,----,---.,---~~____,
,-.,
1.2
OJ
0)
en
N---E 0\
~0.9
-
tr)
0
><:
0
0.6
a)
0.3
0 40 80 120 160 200
8 (degrees)
2.0 , , , 1
Sugar maple
1
~ ;-
-6 1.2 r
0
o
1
.6
J~
;' t= T
y
x F
('
;kt
__
\ .....
X
'y
- ,,/
--- +x
G
-x
a)
!r~O'
I
,~ ,' ,
I,' 10,I, Il'r I'ru liT'
1' 0
, ' , i~ ' ~i • t
A 0 D l' H 21' 31' 41' 51' 61' 71'
l ~Jll:-~--Ir---'-----
o~,4 T 2T 3T 4T 71'
• t
B F J
900 x -~ -180 0 y-1' -1800 y-1' -180 0 y b)
the applied fields are never completely homogeneous and therefore result in
dephasy, giving an induced signal recovery as e( - tlT *, ) where T* 2 depends on the
homogeneity of the field.
The free induction decay (FID) curve describes the decay of the induced
signal arising from free precession of the nuclei in the field Bo. The Fourier
228 Nuclear Magnetic Resonance
en
~~~~~--+---+---+---~--~--~---r--~§
~.~~--~~~--+---+---~--~--~---r--~1 .-5~
~~~~~~~=r~~~~~~~I~ , 0)
~~--~~--4-~~~~~-+--+-~I~
I
r- 1"0..
: ~
. . . . . . . . . . . . . . . . . . .time
2 3 4 5 6 7 8 9 x SOO/lS
Fig. 6.9. Free induction decay (FID) amplitude of 'H nuclei in maple with different moisture
content as a function of time following a 90° pulse. Note that the signal increases in amplitude
and the relaxation rate decreases with moisture content. (Sharp et al. 1978; courtesy of Wood and
Fiber Science)
1.0
.el
'\::
::l
.rj
ta
"@
.§
en
0.5
Q
~
.......
0 o maple
"0
(1)
• spruce
::l
.'=
0..
§
Fig. 6.10. Amplitude of FID signal measured at 50~s as a function of moisture content. (Sharp
et al. 1978; courtesy of Wood and Fiber Science)
.~
"@ - - - - - -
signal from wood and water 54% M.e.,
o
~ \ ~ signal from cellular wall 0% M.C.
Fig.6.11. Free induction decay at three moisture content levels. (Hailey et al. 1985; courtesy of
Forest Products Society)
the wood structure with a superimposed additional signal from the absorbed
water. No mobile protons were detected on the signal corresponding to 0%
moisture content. The mobile protons of water in a sample of western red cedar
of 25% moisture content can be quantified on the graph as the intersection of
the profile of the mobile protons with the axis of the signal amplitude.
Menon et al.(l987) measured the IH NMR free induction signal for western
red cedar sapwood at three moisture contents: 216,92 and 54%, and observed
three relaxation times (Fig. 6.12). The very rapid component at 54% is related
to Tl and is <lO/-1s. Two components were measured for T2, a fast one, decay-
ing at T2 fast < 100/-1s and a slow one decaying at T2 slow < 200/-1s. The intensities
230 Nuclear Magnetic Resonance
lOIr-------------------------------~
'-" "
Western Red Cedar
\.
p ....
'~lOo \. ....
<J) •
.5 "" '216%
tl '.
:::s
~
u 92%
WI
y 54%
I I I I
Douglas Fir
., •• '99%
47%
Fig.6.12a,b. CPMG decay curves. a Western red cedar at 216, 92, 54% moisture contents. b
Douglas fir at 99 and 47% moisture contents. N.B. The logarithmic intensity scale that could give
a straight line for a single component exponential decay. (Menon et al.1987,Academic Press, with
permission)
of CPMG data normalized versus the solid proton signal from FID at each
moisture content were plotted versus time for western red cedar and for
Douglas fir. The authors detected three components of the relaxation time on
visual inspection of the slope of the regression line. The spin lattice relaxation
curves are decaying nonexponentially except at the highest moisture content.
It should be noted that the protons near the surface interact with the cell walls.
T2 becomes a measure of pore size.
The spin-lattice relaxation time T J can be determined from an inversion
sequence illustrated in Fig. 6.13. Immediately following the magnetization a
180 0 pulse is reversed. The magnetization then reverts to its equilibrium value
as Mz(t) = Mo[1-2 eH/TIl ]. This defines what is known as the longitudinal relax-
Basic Aspects of NMR Imaging 231
z z z
a) +~
~Y 1 - - -- Y A-"!,- - -Y
\ :
\'
x t=O x x J
A C M0
z z
_."MO
/ "1
, Y
..r--"t-- Y
\ I
x ~ x
D Mo F
b)
I
'" 180° x
1!~ ';
~~c:%~·~--~-
I
90 0 -x
"....,"">"l;me
------=--------'
~r~--~r~----~---
~ I'" I"'",
180° x
~),~·~--~~~II~'m-~
90 0 -x
A E A E
Fig. 6. 13a,b. Inversion recovery pulse sequence and NMR signal for TI measurement. a Inversion
recovery spin isochromatic representation. b Pulse sequence. (Hailey et al. 1985; courtesy of Forest
Products Society)
ation time T 1• This time can range from ms in metals to several hours in insu-
lators at low temperature and is determined by local fluctuation in the field
experienced by the individual nuclei but at the Larmor frequency. This decay
depends on the nuclei concerned with larger fluctuations in a liquid than in a
solid. For example, the nuclei in the water of wood comes into equilibrium
faster than nuclei associated with protons in the cellular structure.
Observation of the amplitude immediately following a subsequent 90° pulse,
enables T 1 to be measured and is convenient for varying the recovery time 'to
This assumes that at the start of the sequence M z(t) equals M equilibrium, but
if T 1 is very long this may not be true. To make measurements on a specific
nuclei subsystem, the time between pulse sequences 't must be longer than the
232 Nuclear Magnetic Resonance
time Tl (Tl > "[). If the time between pulse sequences is shorter than the relax-
ation time ("[ < T 1), no signal from the nuclei is observed.
For the measurement of T2, more sophisticated pulse sequences (90°, 180°,
180°, etc.) than for TI are used. The presence of the component relaxation times
T I and T2 is attributed to the different mobility of water molecules in different
structural regions, as for example the free water in the lumen and the bound
water in the cell wall. Molecular processes and the total signal intensity depend
on these two times as well as on spin density and proton mobility.
Since the FID signal is proportional to proton density, the moisture content,
MC, in wood can be deduced using
(6.5)
where MC is the moisture content of wood measured with the NMR technique;
PHwood is the proton density in wood calculated from the amount of cellulose
and lignin ; PH water is the proton density in water; Mo is the zero intercept of
the FID curve of the mobile component So is the zero intercept of FID of the
total signal component. The proton density is expressed as the number of
protons per gram. The ratio PHwood/PH water is 0.549 for western red cedar and
0.538 for Douglas fir.
The agreement between the NMR moisture content measurements [MC]
and routine oven-dried moisture measurements [MC] has been shown with
the regression equation calculated for the sapwood and the heartwood of
western red cedar and of Douglas fir, which is R2 = 0.9916
MC = 0.9791 MC-0.7561 (6.6)
This agreement is a milestone in the further application of NMR techniques
in wood science since it demonstrates that it is possible to measure directly the
moisture content of wood without oven-drying the sample. From such data an
upper limit on the difference in moisture content between the heartwood and
the sapwood of the same species can be established.
Because nuclei in different environments decay with different characteris-
tic time constants, Olek et al. (1994) used NMR protons to distinguish bound
water from free water in birch. The spin-spin relaxation time T2 was measured
with a pulse spectrometer at 30MHz and with the CPMG method. The time
between It pulses was 200).is for samples with a moisture contents close to the
fiber saturation point and 5000).is for samples with an even higher moisture
content. The repetition time between pulse sequences was 5 s. The multi-
exponential spin echo decays were calculated as a function of a two-
component relaxation functions, such as:
A(t) =AAexp(-t/T2A )+A Bexp(-t/T2B) (6.7)
where A(t) is the amplitude of spin echo at time t; T2A, T2B are the spin-spin
relaxation times of the spin subsystems and A Aand AB their amplitude con-
tributions (abundances) of the subsystems.
Basic Aspects of NMR Imaging 233
10~~~~~~~~~~~~~~~
8 77 % moisture content
:> 6 "
o
~4 0 00
.€
-a 00
00
§2 o
000
o 0
o
1
0 10 20 30 40 50 60 70 80 09
a) time (ms)
6.0
0
4.0 0 10% moisture content
o 0
:> 2.0 0
OJ 080
~ b:g ao
i
0
o 0
O.6 0
0.4 000
0.2 0
0.1
0 2 3 4
b) time (ms)
Fig. 6. 14a,b. Spin-echo amplitudes versus time in birch above (a 77%) and below (b 10.4%) the
fiber saturation point. (Olek et al. 1994, with permission)
The spin echo amplitudes versus time for a saturated specimen (77%) and
dry specimen (10%) are shown in Fig. 6.14. The spin echo's decay with time is
in the range 0-90 ms. The sample with 77% moisture content clearly shows two
distinct relaxation components corresponding to T2A, 60 ms and to T2B, 100 ms.
For the dry specimen the decay is also non exponential and characterized by a
relaxation time «4ms).
The moisture content MC above and below the fiber saturation point can be
expressed as a function of the corresponding relaxation time T2 (respectively,
T2B and T2A ) and is then calculated with:
T2 = MC/(a+ b MC) (6.8)
where a and b are empirical constants determined for birch as: a = 2.481 and
b = -0.0124. above the fiber saturation point and a = 18.67 and b = -0.1503
below the fiber saturation point and MC is the gravimetric moisture content
of wood, determined by routine measurements.
The relationships between the abundances AB and AA and the moisture
content above the fiber saturation point are given in Fig. 6.15. The abundance
234 Nuclear Magnetic Resonance
100 <=:C",~~;-r-~;-;-~~;-r-;-r-r,---,~,~,~,~,,-,'~
90f 0 j
:
80[ 8 ~
~ ~gt
0..:r: 50 _ o
..:r: 40 AA =-1.281·MC+136.9 o
o
30
30 40 50 60 70 80 90
a) moisture content (%)
30 40 50 60 70 80 90
b) moisture content (%)
Fig. 6. 1Sa,b. The abundance AA of the subsystem A and the abundance AB of the subsystem B vs.
the moisture content in the range 30 to 90% moisture content. a AA of the subsystem A vs. mois-
ture content. b AB of the subsystem B vs. moisture content. (Olek et at. 1994, with permission)
14~~~~~~~~~~~~~~
12 MC o
~1O
T2A 18.67-0.1503'MC
'" 8': r =0.99
S
~ ';
..:;
f-;<'1 6
4
0
30 40 50 60 70 80 90
a) moisture content (%)
70
60 MC 0
T2B 2.481-0.0124·MC (1'0
50 000
r =0.98
40
30-
g'" 20'
~
E-< 10
0
30 40 50 60 70 80 90
b) moisture content (%)
Fig. 6. 16a,b. The spin-spin relaxation time T2A and T28 vs. the moisture content above the fiber
saturation point. a T2A vs. the moisture content. b T28 vs. the moisture content. (Olek et aI. 1994,
with permission)
Table 6.3. Spin-spin relaxation time (ms) expressed by fast, slow and intermediate components
in birch wood as a function of the moisture content. (Data from Olek et aI. 1994, with
permission)
T2B and varies in the range 2 to 17 ms and also increases with the moisture
content above the fiber saturation point (Table 6.3). The low spin-spin relax-
ation time T2A indicates that the very low mobility of the water molecules asso-
ciated with this system, is mainly related to bound water.
Below the fiber saturation point, we have seen that the decay of spin echoes
is nonexponential. Two relaxation times were measured, the first T2A, ranging
between 0.4 and 4 ms and a relatively slow time called T2C, ranging from 2 to
10ms. This time is shorter than the time involving capillary water. Both T2A
(Fig. 6.17) and T2C (Fig. 6.18) increase with the moisture content. For the rela-
236 Nuclear Magnetic Resonance
4.4 ,-r-rrl,,--r---..,,--.-r.--r,,'-'.,---rrl-.--r---..-'-'-,
4.0:" MC
3.6 . T2A l4.76-0.2544.MC
~3.2. r =0.87
S 2.8 :-
<2.4
(',12.0
E-< 1.6 a
0 00
1.2 o
0.8
o
0.4
0.00Ld::.J-"-L--L5-'-'-.L-Ll'-0L.L.J-"-L15-'-'-.l....L2'-0'-'-L..L..1
2 5.-L-L--L..l....J
30
moisture content (%)
Fig. 6.17. The spin-spin relaxation time T'A vs. the moisture content below the fiber saturation
point. (Olek et al. 1994, with permission)
16
14 MC
T 2C
12 4.9l5-0.1009·MC
~1O r=0.81
8
u8
~6 co
4 0
0 OJ
2
00 5 10 15 20 25 30
moisture content (%)
Fig. 6.18. The spin-spin relaxation time T,C vs. the moisture content below the fiber saturation
point. (Olek et al. 1994, with permission)
In this section, we will outline the NMR imaging technique, with special stress
on what the images represent and on the algorithms used to identify and isolate
particular structural features of wood.
The magnetic resonance technique can image nuclear spin density, NMR relax-
ation times and diffusion of water molecules at various points throughout the
wood sample. The values of relaxation time T1 and T2 are important in deter-
mining the quality of an image and can be used in selective imaging of bound
and free water molecules in a wood sample under certain conditions. The
intensity can be proportional to the product of the proton density and the ratio
of the relaxation times. High resolution NMR imaging is very sensitive to the
more mobile protons. The dependence of the image intensity on relaxation
times is very important in differentiating between wood tissues with the same
proton density but different values of the ratio of the relaxation times. In
general, the quality of an image depends on many parameters. It is also possi-
ble to vary the experimental conditions in such a way that the image is par-
ticularly sensitive to a specific chosen parameter. The spatial discrimination is
dependent on the magnetic gradient field.
The techniques currently used in NMR image reconstruction are: back pro-
jection, similar techniques used in X-ray computed tomography, rotational
frame, echo planar for two-dimensional images and two-dimensional Fourier
transform techniques for three-dimensional imaging of the object. The Fourier
transform technique, which encodes the spatial information in phase with the
NMR signal, is optimal with respect to signal-to-noise ratio. The imaging of
porous solids like wood requires particular considerations. The solid matrix
material has an important effect on the relaxation, which can limit the resolu-
tion of standard imaging methods. The different water phases in wood must
be observed separately if possible. If not, the proportion of measured signals
corresponding to various phases must be distinguished.
For all imaging methods, there is a strong dependence of imaging time on
spatial resolution. The imaging time depends on spatial resolution as (resolu-
tion)4. However, the spatial resolution tends to decrease as the size of the object
increases. Typically, resolution is currently of the order of 1-2mm. In the
future, improvements of the technology will be directly related to the improve-
ment of acquisition time. The usual approach to obtain 3-D images is to use a
2-D image of a selected thin slice of a 3-D object, and then to move to the next
slice and so on until the whole object is reconstructed A thin slice is defined
by using frequency-selective radio frequency pulses, or a time-dependent field
238 Nuclear Magnetic Resonance
a)
b)
Fig.6.19a,b. Outline of the NMR system. a Block diagram. b Position of the log in the center of
the magnetic field. (Wang et al. 1990, with permission)
Fig. 6.20. Three NMR imaging views of a log. a Coronal view, from top to bottom. b Sagittal view,
from left to right. c Transverse view. (Chang et al. 1989; courtesy of Forest Products J)
a) b) c)
Fig.6.21a-c. Electric current density imaging of a young oak tree. a Cross section in light
microscopy. b Nuclear magnetic resonance image. c CDI (current density image). A Cambial zone
+ differentiating xylem + complete phloem; B differentiated secondary xylem of the current
growth ring; C first annual ring; D pith. (Beravs et al. 1998; courtesy of Holzforschung)
)-r:,;, ho'~
x
, /
;0'01
RFcoll
- wood
_'''-.. glass tube
RF input/output
Fig. 6.22. The NMR probe containing the wood sample. (Quick et a1.l990; courtesy of Wood Fiber
Science)
6.3.2 Algorithms
I:: I::
0.0 0.10 0.20 0.30 0.40 2 .....;::l
time (s) U
'l)
U
'l)
0..0..
en en
10000
0.006: '
8000 I
'l) I
I
]6000 I
I
~OOO I
§2000 b I
I
I
0
10- 3 10-2 10- 1
T2(s)
Fig.6.23a,b. Linear inverse technique to characterize decay curves in terms of discrete or con-
tinuous distributions of relaxation times. a Amplitude vs. time. b Amplitude vs. T2 with discrete
and smooth spectra. (Whittall and MacKay 1989, Academic Press, with permission)
242 Nuclear Magnetic Resonance
For simple systems involving protons in water, the relaxation of Mz and Mxy are
given by simple exponentials with relaxation times T1 and T2' However, wood
is a naturally occurring composite material, where the water molecules in dif-
ferent macroscopic regions have very different relaxation times, where the
values of relaxation times from different regions are significantly different.
In some circumstances it is possible to extract T 1 and T2 for the protons in
different regions of the sample by fitting the relaxation to two exponential
terms. Some authors (Menon et al. 1987) have even attempted to describe the
relaxation process with three separate relaxation times. However, to obtain reli-
able information in such a way with any confidence would require an extremely
high signal to noise ratio and very long relaxation times to enable the various
regimes to be differentiated. For example, with an observation value for T2 of
6 and 200 ms for water in the wood, a wide range of permeability measure-
ments is possible for small and large anatomic elements (fibers, vessels, etc.)
which are in fact capillary tubes.
Another important feature of NMR imaging is its ability to identify wood
structure and defects. An example of this is given by the pulsed NMR mea-
surements of black oak (Quercus velutina) samples used by Coates et al. (1998).
An example of the identification of major structural areas is shown in Fig. 6.25.
The spatial variations observed are probably due to variations of the moisture
content in the sample and associated variations of density in growth rings and
rays.
In all defective regions, the gray level of some pixels widely deviate above
and below the gray levels of the clear wood region. In clear wood areas, the
variations in the gray levels are also important due to growth rings and rays
also because of random noise. Constant threshold limits are not able to give
the degree of variability of gray levels found in a particular log. A more effec-
tive approach is to introduce threshold limits that follow the contour of the
average moisture content. The largest gray level deviations are associated with
the defect area. Using values obtained for clear wood zones and subtracting
the largest deviations associated with defect areas, it is possible to obtain a
clear map of the defects. Computer algorithms have been developed by Vincent
(1993) and Peters (1995) to enhance the imaging of defect regions (Tsai 1985).
The NMR Imaging Technique 243
Transversal section
Fig. 6.24. Imaging of transverse and longitudinal radial section of a beech. (Photo Escanye, in
Bucur 1990, with permission)
244 Nuclear Magnetic Resonance
100
80
/
pith
\ knot
scar
40
20
~'I
0.0)0 0.100 0.1)0 0200 0.250
X lIXi$ ( 10e3)
d)
Fig.6.2Sa-d. Comparison between the photographic image and the automatically segmented
image of oak, using the algorithm proposed by Coates et al. (1998). a Photographic image of the
cross section. b Magnetic resonance image of the section. c Deviation of 40 gray levels or more
between original pixel and median axis. d Central profile of c where pith, knot and scar are visible.
(Courtesy of Forest Products J)
6.4 Applications
In this section, the physical and chemical characterization of wood using NMR
imaging techniques are discussed in relation to the inspection of trees, logs
and lumber and the possible industrial control of drying, impregnation of
wood products with preservatives and adhesion. For all such applications, it is
important to be able to quantify the water content, and the corresponding
Applications 245
moisture content in wood, and, on the other hand, the structural features at
macroscopic and microscopic level. Different aspects of wood processing
include the control of the moisture content in lumber during drying, the
control of the moisture content in wood-based composites, the control of
impregnation processes with preservatives and the kinetics of curing during
adhesion.
NMR imaging can also be used to image bulk physical properties of wood,
such as electrical conductivity. For many applications, it would also be useful
to be able to image the elastic properties of wood in situ. A very interesting
field of potential application is the simultaneous measurement of elastic con-
stants of bulk objects. Until now, this approach has not been used for wood
materials. Nevertheless, we present a theoretical background for this applica-
tion, which seems to be very promising for the mechanical quality estimation
of green lumber, trees and logs.
Examples of NMR imaging for the in situ inspection of living plants have been
given by Halloin et al. (1994) and Nakanishi and Matsubayashi (1997). Exam-
ples illustrating distribution and movement of water in roots and meristems
during development, in leaves or needles, and seeds have been reported.
Mac Fall et al. (1990) investigated the water depletion region in loblolly pine
roots. During the first stages of growth, the roots were easily visible and the
formation of a distinct water-depletion region around the roots was evident.
Later, this region extended around the lateral roots and clusters of mycorrhizal
short roots. The development of roots of Douglas fir seedlings, grown in con-
tainers at normal temperature, has also been investigated (Brown et al.I991).
The lengths of the roots determined from NMR images was in agreement with
the measured lengths of excavated roots. Southon et al. (1992) studied the influ-
ence of freezing on the roots of Scots pine and Norway spruce seedlings over
a large temperature range from +lto -20°e. The amount of root material
detectable by NMR imaging decreased with decreasing temperature, presum-
ably because of a significant increase in T1 and T2 in the material as the polar
mobility of water molecules, particularly in root material, decreases with the
temperature.
Several aspects of leaf and needle anatomy and morphology have been
discussed by Masuch et al. (1991), Veres et al. (1993) and Millard and Chudek
(1993). In vivo measurements of growth were made as a function of the water
content of the soil. The NMR signals of needles from declining forests showed
a larger NMR signal and a different distribution of free water from that of
healthy needles. NMR imaging has been also used for the discrimination of the
germinability of seeds as noted by Vozzo et al (1996).
246 Nuclear Magnetic Resonance
In this section, we will discuss techniques for water content imaging in living
trees, in fresh logs and in green timber. Previously, it was noted that the nature
of NMR signals is strongly dependent on the amount of water present in dif-
ferent wood tissues. The presence of magnetic ions (Cu, Fe, Mn) in the sap and
in the cell walls causes a strong reduction in the nuclear relaxation times T1
and T2' For example at 1% of copper sulfate solution in water reduces T1 and
T2 from about 3 s to 20 ms. NMR imaging and spectroscopy can also provide
information about the anatomic structure of healthy and infected wood of
living trees which will be discussed in this section.
It is generally agreed that increment cores bored into the trunk of living trees
are representative of the quality of the tree. Byrne et al. (1986) pioneered the
measurement of the water content of trees by using an NMR spectroscopic
technique on 5-mm-diameter increment cores bored from healthy young Pinus
radiata and Eucaliptus grandis trees, using NMR at 60 and 90 MHz at two tem-
peratures (20 and 30°C), to investigate their frequency and temperature depen-
dence, as a noninvasive technique for measurement of water flow in young
trees. NMR measurements of T1 and T2 were made using two spectrometers.
T1 was measured by the 180°,90° sequence inversion recovery pulse sequence.
T2 was measured from the decay of spin-echo signals as a function of pulse
spacing in both species. The two measured relaxation times T1 and T2 had two
constituents, one long and one short, and were determined from logarithmic
plots of the NMR decay envelope at ambient pressure and temperature. The
moisture content of the two specimens ranged from 39 to 43%. To identify any
changes between the freshly taken samples and those used for laboratory
experiments, specimens were tested in their fresh state and after 3-day storage
in a sealed container.
The results are given in Table 6.4. The scarcity of experimental data permits
only quantitative conclusions:
The data reported in this paper suffer from some inconsistencies (reference
to fresh and old eucalyptus samples). The presence of the two-component
relaxation times can be attributed to the different mobility of water molecules
Applications 247
Table 6.4. Some values of the relaxation times T, and T2 (ms) measured for Pinus radiata and
Eucalyptus grandis at 40% moisture content (Byrne et al. 1986,2002, with permission of Elsevier
Science from Agricultural and Forest Meteorology, vol. 38: 307-317, Bynne GF, Fenn MD, Burgar
MI 1986, reprinted from Nuclear magnetic resonance studies of water in tree section)
Table 6.5. Average values of electric current density, of electric conductivity as a function of
wood density and wood moisture content in a young oak tree .(Beravs et al. 1998, courtesy of
Holzforschung)
pixel grid 256 x 256. The imaging time was 15 min to average six scans. The
voltage applied to the sample was 180V, with a current pulse length of lOms.
The 3-D images were obtained from two perpendicular orientations of the
sample. As is evident from the image, current density imaging offers a higher
contrast between the different tissues than conventional NMI imaging.
The current density image shows a conductivity contrast in four different
regions of the oak sample. The corresponding values of electric current den-
sities, electric conductivity, and density and moisture content of the wood are
shown in Table 6.5. The increase in the moisture content from the first annual
ring to the cambial zone results in an increase in all measured parameters.
All reported experimental data in this section were obtained at room tem-
perature. The influence of low temperatures on relaxation times was studied
qualitatively by Johansson (1985) with frozen logs in Sweden. No NMR signals
were obtained in this case, probably because of the solid state of frozen water.
The mobility of water molecules decreases with decreasing temperature so that
there are fewer higher frequency components of the local fluctuating magnetic
fields to restore Mz to its equilibrium value (hence TI increases) with similar
effects on T2'
The second part of this section is devoted to the imaging of fungal coloniza-
tion of young living trees. The imaging of fungal attacks with different patho-
logical agents of living sapwood of sycamore has been reported by Pearce et
al. (1994, 1997). The infection agents for sooty bark disease were Ustulina
deusta, Chondrostereum purpurem, Cryptostroma corticale, and Ganoderma
adspersum. The development of fungal lesions in the sapwood of living trees
is restricted by an active defense mechanism. The pathological anatomy of
sooty bark disease lesions can easily be seen by imaging shown in Fig. 6.26.
The inoculated zone is surrounded by a reaction zone located between the
Fig.6.26a,b. Different sections through a stem of young Acer pseudoplatanus tree wound inocu-
lated with Ustulina deusta. Longitudinal-radial section inoculated with Ustulina deusta where
one can see D decayed zone; H healthy zone; R reaction zone, which separates the healthy and
decayed zones; C callus beginning to grow over the wound faces. Scale bar 10 mm. a NMI obtained
with a spin-echo image sequence for a 2-mm slice thickness, with a resolution higher than
0.47 mm. b Comparison of pixel intensities in specific areas of Mo maps and maps of measured
TI and T, have shown that both techniques are similar for the illustration of different lesions (L)
in different tree regions. Reaction zone had a high contrast. Scale bar 10mm. (Pearce et al. 1997,
Blackwell Science LTD, with permission)
250 Nuclear Magnetic Resonance
decayed and the healthy wood. It was demonstrated that changes in moisture
content are associated with the formation of decay lesions in living trees. Spin-
lattice Tj maps showed little contrast compared with proton density maps.
Measured relaxation times were within the range of 700 to 1000 ms.
The spin-spin relaxation time T2 for healthy wood varied between 35 and
127ms. In reaction wood, T2 was reduced by as much as 40%. Neither Tj nor
T2 was altered in the drier wood adjacent to naturally occurring Cryptostroma
corticale, but both were reduced in the infected wood. The moisture content
was higher in the reaction zone than in the healthy area. Calculated proton
density images showed the delimitation of the reaction zones. The contrast of
the images of lesions was attributed to proton density variations and to the
decrease in spin-spin relaxation time T2 in the vicinity of the lesions. In the
reaction zone, the values of spin-lattice relaxation time T were less sensitive
j
6.4.2 Growth Rate and Other Structural Features in Logs and Lumber
The detection of annual rings, knots, reaction wood, wet wood, decay and other
structural features with the NMR technique is a basic approach and is a chal-
lenging task for image reconstruction and processing. Pattern recognition
techniques to identify, classify and quantify structural features are needed for
the on-line internal inspection of logs before sawing at normal production
speeds in saw mills and other wood processing. NMR image processing tech-
niques are similar to those used for X-ray computed tomography, but the infor-
mation content of the NMR images is higher than that obtained from X-ray
computed tomography and ultrasonic imaging. Because the NMR signal inten-
sity is directly proportional to the moisture content, the higher the moisture,
the better the image. However, even for samples with a low moisture content,
a reasonable image can still be obtained by averaging signals over a long time.
Hall et al. (1986a,b), using a relatively low field of 0.14 tesla, obtained clear
images of the macroscopic structure of fresh-cut bolts of aspen and spruce of
25 cm diameter. Annual ring patterns, buried knots and decayed regions were
clearly observed. Similar images have been reported by Wang and Chang
(1986) and Swanson and Hailey (1987).
In this section, we present images from a beech bolt of irregular shape and
elliptical cross section (75cm long and maximum diameter 15cm), cut from a
tree at a height of 20 m from the ground, on which bolt branches were formed.
In this sample, the wood structure is very complex, with normal wood, reac-
tion wood, knots and other features. Beech was chosen for imaging because of
its commercial importance in Europe. The internal structure of the bolt is given
Applications 251
a) b)
c) d)
Fig.6.27a-d. Views of the internal structure of a beech bolt. Images in the transverse plane at
different heights (74, 33,21, 15cm) noted a, b, c, and d, respectively. NB: The images were taken
on a medical scanner for 0.5 tesla, with a spatial resolution of 1.2 x 1.2 mm. The echo time
(30 ms) permitted imaging only of the distribution of mobile, free water. (Photos Escanye in Bucur
1990, with permission)
in Fig. 6.27 with four transverse sections at different heights (15,21,33, 74cm)
and sagittal views. Well-imaged features are the annual rings, the pith, the
red heart zone, the sound wood, the tension wood, and even the bark. The
image clearly shows branch formation, which corresponds to the physical state
of the bolt. The bark is seen as a very clear zone. Loss of image intensity is due
to the low water content. The oval-shaped bright areas correspond to branch
formation with decreased water content and reveal the existence of buried
knots.
252 Nuclear Magnetic Resonance
The abnormally bright area around the pith corresponds to red heart. The
development of red heart is different at different heights. The dark protuber-
ances of the red heart contain less water than the surrounding area, probably
because of to the presence of tyloses.
The diameter of the pith is 3 mm, and the first annual ring has a diameter
of 4 mm. In the healthy zone, the earlywood of the annual rings shows a lighter
color than the latewood of the same ring, which implies either that it contains
more water or that this water has a longer relaxation time. The latewood zone
in all rings is always dark, indicating low water content. The natural distribu-
tion of water during the life of the tree can be seen. The influence of climatic
changes, or of physical and biological damage can be determined from such
NMR images. By looking at the image carefully, one ring of the tension wood
with a darker latewood zone can be seen in transverse image at 25 mm from
the periphery of the bolt. In the sagittal plane image, the same dark vertical
projection of the ring can be observed. This ring is therefore likely to be tension
wood with less moisture content than the surrounding rings and a shorter
relaxation time. White irregular small bright stain checks between the annual
rings signify a high level of water content
It is not the aim of this section to review the NMR techniques that reveal the
porous nature of solids in general, but rather to select the studies on those
materials of most relevance for the understanding of wood. Cohen and
Mendelson (1982) studied the influence of the size and geometry of pores in
rocks on NMR signals. Assuming the validity of the Bloch equations with mag-
netic relaxation on the surface of the pore voids in a uniform field, T 1 and T2
could be defined as a single relaxation rate that was related to the surface area
and the volume of the pore spaces. The NMR measurements of T 1 and T2 were
made to determine the spin-lattice relaxation and the diffusional attenuation,
using a spin -echo signal. A strong magnetic field was used to polarize the water
molecules in the pores of the sedimentary rocks. When the magnetic field is
reduced, the proton magnetization relaxes to a new equilibrium value. The
decay in the amplitude, at a rate determined by T[) is sensitive to the geome-
try of the pore space of the rock. Similar approaches can be extended to the
specific case of porous woods, and a new method for porosity measurement
could be developed in the future.
In this section, we consider the sensitivity of the nuclear magnetic imaging
technique to the structural features of wood on a microscopic scale, as a func-
tion of cell size and the dimensions of earlywood and latewood in the annual
ring.
Wycoff et al. (2000) estimated the cell size in four softwood species (eastern
red cedar, redwood, eastern white pine and sugar pine) using NMR measure-
ments. The diameter of the cells in these species is between 25 and 65/-lm and
Applications 253
longitudinal
2.0
1
'"
0
~ 1.5
.~
IE
§
"
.~
@
'il
0.5
0.0 '---_I....--"I....---l_--1_--L_....:.._-I.._....:.._ _
o 200 400 600 800
Sugar pine 42 ± 3 45 ± 3
Eastern red cedar 17 ± 1 17 ± 2
Eastern white pine 31 ± 1 35 ± 3
Redwood 40 ± 1 43 ± 3
sured for the field gradient in the longitudinal direction are very close to the
values for bulk water (2.2 x 1O-9 m 2/s), demonstrating that there is no signifi-
cant barrier to diffusion in this direction. The deduced diffusion coefficients
were between 2.1 x 10-9 and 2.4 x 1O-9 m 2/s. The diffusion length along the cells
corresponding to a diffusion time of 903 ms corresponded to a diffusion length
of 150 ~m. For pulsed gradient fields in the tangential direction the barrier
walls limit the diffusion length.
Good agreement between the dimensions of cells measured by optical
microscopy and NMR is evident from the data shown in Table 6.6. The values
measured with the NMR diffusion technique, using water molecules as probes,
were only slightly higher than those obtained by microscopy. Wycoff et al.
(2000) demonstrated that the NMR provides accurate measurements for cell
sizes at moisture contents above the fiber saturation point. However, for dry
wood, it would be more difficult to separate the signals for free and bound
water.
One-dimensional NMR microscopy can be used to produce images at the
millimeter and submillimeter level corresponding to regions with different T2
values. Menon et al. (1987) demonstrated that NMR signals from water in
western red cedar (Thuja plicata) can be separated into three components with
different T2 values such as:
- values of a few milliseconds corresponding to water bound within the cell
walls,
- values of about 50 ms corresponding to water located in latewood tracheids
and ray lumina,
- values of over 100 ms corresponding to water in the lumina of the earlywood
tracheids.
It was also found that the relaxation time corresponded to lumen water T2
scales with lumen diameter (D) as D2.
The image in Fig. 6.29 was obtained with a Carr-Purcell-Meiboom-Gill
(CPMG) pulse sequence in a constant field gradient of 19.4 Gtesla/cm. To
isolate an image of each T2 component of water in wood, it is assumed that
Applications 255
Fig. 6.29. Images of free water in the lumina of earlywood tracheids (EW), of the lumina of
latewood tracheids (LW), and from bound water in the cell wall (CW) superimposed by the
scanning electron micrograph of western red cedar. (Menon et al. 1989, Academic Press, with
permission)
where t is the time (ms) from the start of the CPMG sequence and S(I) is the
amplitude of the CPMG echo sequence.
The three terms of the equation correspond to the three T2 components of
water content. When the water content in latewood, earlywood and cell wall
components were superimposed on the corresponding scanning electron
micrograph, near perfect agreement was observed. The authors also demon-
strated that the variation of bound water obtained from one-dimensional
images across the annual rings on a green specimen is perfectly matched by
that of the oven-dry wood density in the radial direction when measured with
an X-ray micro densitometric technique on the same specimen, as can be seen
from Fig. 6.30. The major contribution of Menon et al. (l989) has been to
demonstrate that NMR imaging can be used to separate images on the basis of
spin-spin relaxation time T2, so that images on a submillimeter scale can be
achieved. Such studies will enable the role of the bound water to be better cor-
related with other physical properties, such as dimensional stability, thermal
and electrical conductivity and mechanical properties.
256 Nuclear Magnetic Resonance
1000
400
800
1600
.€
'"
~ 400
"8
~ 200
o 2 3 4 (rom)
Fig. 6.30. The profile of the bound water in the cell wall compared with the X-ray densitometric
profile of annual rings in western red cedar. (Menon et al.1989,Academic Press, with permission)
, , ,,
I
td 90°:, ts :90°
,
tm i
90 0 I tc I,, ta
,
I
..
I I
I ~ I
+x: +y : ±x +x~ +y :
I , '----
I
I ,
Fig.6.31. Representation of the pulse sequence MOPS to monitor [H and BC spin diffusion.
t, Contact time; t. data acquisition time; td recovery delay time; t, spin locked time of 8 ms; tm
remaining magnetization time; -x and +x pulses phased along x-axis. (Newman 1992; courtesy
of Holzforschung)
Because the Larmor frequencies are specific to different nuclei, NMR methods
can be used to investigate the variation of the chemical composition of a wood
sample. The spatial variation of different chemical components in the wood
cell wall was first studied by Newman (1992) using NMR. He found that natural
differences in the proton spin relaxation time constants governed different T2
values corresponding to ordered and less ordered cellulose and amorphous
lignin. The cellulose signals with chemical shift 0 = 89 ppm were assigned to
13C in cellulose and 0 = 56 ppm 13C at nuclei in lignin. T1 and T2 for 13C can be
obtained from the same experiment at 50.3 MHz as described in Fig. 6.31 for
Applications 257
0.9
'-'
en
.9
.=:
v
0.8
u
6\
00
~
a.
::I:_
f--; 0.7
.S
c '1
1f ""1------1 Eucalyptus
\, , ,....l
<;:;
'-D
~
, ,
0.6 ,,~~
a.
:r::_
f--;
0.5
0 20 40 60 %
water content (%w/w)
Fig. 6.32. Proton rotating frame spin relaxation time signals vs. the water content for E.
delegatensis and Pinus radiata and for cellulose and lignin. (Newman 1992; courtesy of
Holzforschung)
Eucalyptus delegatensis and Pinus radiata specimens. The spin locked time was
ts = 8 ms. The value of tm, the time needed for the remaining magnetization
to return to an orientation parallel to the static field, was variable. Proton
magnetization was monitored as a function of tm, with a standard cross-
polarization sequence. During the time tm, the phase of the second 90° pulse
in the MOPS sequence was alternated to suppress interference from TI relax-
ation. An important experimental detail to note is that D2 0 was used in mois-
turizing the specimen (with 35% by weight), which results in a longer value of
TI. Figure 6.32 shows the variation of the ratio of lignin to cellulose relaxation
times TIp H in Eucalyptus delegatensis and Pinus radiata versus water content.
Up to 20% of water content, the curves are superimposed. However, above 30%,
corresponding to the fiber saturation point, the values for eucalyptus are
slightly higher than those for pine. The differences between the spin-lattice
relaxation time, TIH , in cellulose and lignin is again evidenced in Fig. 6.33. The
increase in TIp H suggests that the spin-lattice relaxation is dominated by mech-
anisms involving water molecules or hydroxyl groups. The influence of mois-
turizing on Pinus radiata is more important than on Eucalyptus delegatensis
and can be explained by the difference in strength of bonding of water mole-
cules to the different chemical components of each wood species. The average
258 Nuclear Magnetic Resonance
1.0 legend
E = Eucaliptus
P = Pinus
0.4
0.2
'U;'
0.1
;:2::
f-< 0.08
0.06
0.04
0.02
0 20 40 60 %
water content (%w/w)
Fig. 6.33. Proton spin-lattice relaxation time vs. water content for cellulose and lignin in
eucalyptus and pine. (Newman 1992; courtesy of Holzforschung)
Table 6.7. Proton rotating-frame spin relaxation times TlpH (ms) for Pinus radiata moisturized
with H,O and with D,O to 35% of total weight. (Newman 1992, courtesy of Holzforschung)
difference between cellulose and lignin is 22%. To enhance this difference, the
increase in TJpH was demonstrated for Pinus radiata at 35% moisture content,
with the H 20 replaced by D20, as can be seen in Table 6.7.
The experimental values were influenced by diffusion between domains.
Theoretical values were deduced by computing simulation, using the model
cited before with which spatial segregation of chemical components was
observed, as can be seen in Fig. 6.34, where the signal height differences
between cellulose and lignin were plotted versus the time tm. The perturba-
tion of initial magnetization disappeared after about 20 ms of mixing and no
difference between cellulose and lignin was observed. The difference between
the theoretical and the experimental values for cellulose are in the range of
experimental uncertainties, but for lignin these differences are affected by a
Applications 259
This section describes the potential use of nuclear magnetic imaging in wood
processing, and particularly its applications to the control of the moisture
content during lumber drying, quality control of wood-based composites and
their moisture content at different phases of processing, control of adhesion
between adhesives and wood, and the control of impregnation processing of
wood with preservatives. The last part of this section is devoted to a study of
archeological wood.
Over the last 30 years, extensive literature has been published relating NMR
spectroscopy to the measurement of the moisture content in wood. For the
260 Nuclear Magnetic Resonance
forestry industries and for wood science, the most important application of
NMR is undoubtedly the measurement of water flow and diffusion in wood,
both in situ and in vivo specimens. As noted by Nanassy (1973), the physicists
Shaw, Elsken and Rubin in the 1950s were the first to draw attention to the pos-
sibility of using NMR to measure accurately the moisture content of wood.
A serious limitation of this technique, at the time, was the large expense of
the apparatus required. Using a laboratory spectrometer, Nanassy (1973) pro-
posed a graphic method for measurement of the relative intensity distribution
of the narrow component spectral absorption. This approach permitted the
moisture content in yellow birch and white spruce to be derived in the range
of 5 to 100%, on specimens of about 2cm3 volume. The accuracy was around
±3%.A year later, the same author (Nanassy 1974) used a broad-spectrum com-
ponent system to determine the fiber saturation point, called by the author
"critical moisture content" defined in relation to "the phases of the sorbed
water, and the ratio of wood protons participating in chemical exchange
between wood and water protons." The fiber saturation points at 25°C for white
birch and white spruce were both found to be 38% of the protons in specimens.
This value was higher than the generally accepted value of about 28% mois-
ture content for wood species from temperate zones. Nanassy (1974) explained
that the diffusion of water in wood over the cell wall surfaces is dependent on
the bonding energy of the water molecules to the cell walls, and that such water
is "strongly bound to the rigid lattice of wood." The additional free water, called
by the author "mobile water;' is only weakly bonded to the cell walls. Experi-
ments using improved electronics and corresponding NMR systems during the
last 10 years have validated Nanassy's statements, as shown in many articles
published in recent years.
As seen in Section 6.2, the diffusion of water in wood is anisotropic
(MacGregor et al. 1983). Measurements have been performed with two species,
Douglas fir and sugar maple, using a pulsed gradient NMR technique. These
measurements confirmed that diffusion along the fibers is higher than diffu-
sion in the transverse directions (Table 6.8). Water diffusion in maple is more
Amplitude
20 xl 04. - - - - - - - - - - - - - - - - - - - - - ,
o 10
------ u=25%
-----u=20%
-----u=15%
------- u=11 %
-O.5xlO 4i - - - - - . - - - - - - , - - - - r - - - - - - , - - - - - - - i
~~
Fig. 6.35. One-dimensional imaging with proton gradient during drying of small, clear pine
specimens from 25 to 11 % moisture contents. (Lindgren 1994, with permission)
262 Nuclear Magnetic Resonance
They represent the radial moisture profile with amplitudes proportional to the
spatial variation of the water content. For the small, clear specimens an accu-
rate analysis of water distribution above and below the fiber saturation point
can be made.
Since the aim of this section is to relate the 1-D distribution of water protons
and the structural features of wood, it was therefore natural to superimpose
the SEM optical image of the corresponding anatomic structure on the proton
profile (Fig. 6.36). To illustrate this statement, we refer to the results on white
spruce published by Araujo et al. (1992) and to the definition of moisture
content proton density in Eq. (6.6). The drying of white spruce sapwood
from 100 to 17% moisture content proton density was followed in four steps,
corresponding to 100, 86, 59 and 17%. The radial moisture density profile
for total water content, for free water and for bound water was obtained and
superimposed on the SEM image. One-dimensional images were obtained
from the Fourier transform of the T2 signals. They showed that the total area
under the profile of the 1-D image is proportional to the mass of water in the
sample.
One-dimensional images for the representation of the amount of total water
(Fig. 6.36a) can be analyzed for each moisture density level as:
- for 100%, the moisture density is higher in the earlywood zone of all rings,
because of the larger tracheid lumina in earlywood.
- for 86 and 59%, the rate of drying is more important in earlywood than in
latewood.
- for 17%, the distribution of water is quasi-uniform in earlywood and in
latewood. It was stated that this curve corresponds to the bound water.
From Fig. 6.36b, representing the 1-D images of the free water present in the
lumina, it can be seen that the profiles are very similar to those discussed pre-
viously. For the 17% water content sample, the central component indicates the
presence of a small amount of water inside the lumina.
In Fig. 6.36c, showing I-D images of bound water, we observe a superim-
position of the bound water profile on the growth ring profile. This allows us
to conclude that above the fiber saturation point the bound water distribution
is in agreement with the density of solid wood, which is higher in latewood
than in earlywood. Below the fiber saturation point, the distribution of the rel-
ative moisture density is uniform in the earlywood and latewood. With such
evidence, it was possible to suggest that the bound water profile was unrelated
to the wood density.
Another possible interpretation of data produced from T2 spectra is to relate
histograms of the lumen radius to that of the proton volume and to the cell
number (Fig. 6.37). Both histograms in white spruce sapwood and in com-
pression wood show similar distributions.
This allows us to conclude that the lumen radius is related to the proton
volume and that the peaks in the continuous spectra can be attributed to the
anatomical location of the water. Different tissues, for example juvenile wood,
0.6
0.5
""'
5 0.4
M
~
c-
' ih
0.3
c
<>
"'" 0.2
~
E
'8 0.1
E
o
- 0.11--- - , -- -- . , -- - - , - - - - - - ,
-4 -2 o 2 4
a) position ( mm )
0.6
0.5
M
E
~ 0.4
c- 0.3
.;;;
c
<>
"'~" 0.2
2
'"
'0 0.1
E
o
-0. 1 +-----,-- - - - - . - - - - - - . - - - - - ,
-4 -2 o 2 4
b) po ilion ( mm )
0.6
0.5
r'l
E
~ 0.4
C- 0.3
.;;;
c
<>
~ 0.2
2
.~ 0. 1
E
o
-0.1 + - - - - - r - - - - - - . - - ----.------,
-4 -2 o 2 4
c) position ( mm )
0.20
..:!) ..:!)
~
u
cells ~
u
cells
'-H 0.15 '-H 0.15
...
0
(!)
...
0
(!)
,D ,D
S S
;:l
;:l
I:i 0.10 I:i 0.10
(!) (!)
.~ :>
ttl '.g
~
... 0.05
~
... 0.05
[ rrr , ~- .'
protons
protons
0.15 0.15
(!) (!)
S S
;:l
..2
"0 0
:> :>
(!)
0.10 (!)
0.10
:> :>
'"§ '.g
~... ~
...
0.05 0.05
o
rfl-1-ri
15
~ I
20
I
25 o 10 15 20 25
5 10 5
lumen radius (Jlm) lumen radius (Jlm)
a) white spruce sapwood b) compression wood in white spruce heartwood
Fig.6.37a,b. Histograms of numbers of cells and of protons as a function of lumen radius in a
white spruce sapwood samples; b white spruce heartwood compression wood. (Arauja et al. 1992;
courtesy of Wood Science Technology)
0
W- 4 10- 3 10-2 10- 1
a) T2 (s)
18
15 -
,\
{\
/' \
-
12 - ",
-
Q) :i ", \
:', \
"0
.~ 9- ",
0..
§ :'/ -.,I
6- :' I :, \
3-
1\ .. / ",I
f-
/!'~
'/,\
/,~,\
' I '
"y', I ".\
0
10-4 10- 3 10- 2 10- 1 10°
b) T2 (s)
Fig,6.38a,b, Continuous spectra T2 in white spruce, a Juvenile wood (solid line) and heartwood
(dashed line); b in rehydrated samples of heartwood compression wood (solid line), heartwood
(long dashed line) and heartwood with incipient decay (short dashed line), (Arauja et al. 1992;
courtesy of Wood Science Technology)
wood tracheids and in ray lumina. The moisture content deduced from NMR
signals was obtained by "multiplying the ratio of water to wood signal inten-
sities by the ratio of the hydrogen content of wood to that of water:' Compared
to values for oven-dried wood, the NMR values are overestimated at higher
moisture contents and underestimated at lower moisture contents by several
266 Nuclear Magnetic Resonance
160.------------------------. 1000,-------------.
~
140 I--'.~\--------------.- .. --
.. ---.s:-.lo,...w----j
1201-__~'------------~-~-=-~··m~e~dwiullim~
......... slow
---··medium
I early wood
'E fast -fast ~
100f--------/-,--,.,.-.-.+;-1
lumena water
~ 1001-----~------------------~
~ 801------·-~·~,,~--------------__j -_.-._..- ~
1
Fig.6.39a,b. Moisture content determined with NMR technique and the relaxation time T2 vs. the
gravimetric moisture content in cedar. a Juvenile wood; b decayed cedar wood. N.B. The mois-
ture content deduced from NMR signals was obtained by "multiplying the ratio of water to wood
signal intensities by the ratio of the hydrogen content of wood to that of water." (Flibotte et al.
1990; courtesy of Wood Fiber Science)
a)
7 days I day
c) d)
15 days 4 days
Fig.6.40a-d. Transverse views of oak lumber (SO-mm thickness) during natural drying. a One-
day drying; b 4-day drying; c 7-day drying; d IS-day after cutting. (Wang et al. 1990; courtesy of
Forest Products Society)
higher moisture content than the surrounding annual ring area. It is evident
that the drying process takes place from the outside to the side of the board.
The kinetics of drying can be observed for all specimens and information
about the moisture gradient and distribution can be deduced. A 25-mm-thick
board dried at a faster rate than the 50-mm board because it simply takes
longer for water to diffuse out from the center of the thicker boards. The dark
area is dependent on orientation of the growth ring. The brighter area in the
middle region of the cores indicates that diffusion is faster from the central
core areas than from the outer surfaces of the board.
We have shown that NMR imaging is a valuable nondestructive and non-
contact method for qualitative and quantitative studies of the kinetics of wood
drying. Its only drawback is the cost of such equipment, which is similar to the
medical device used in hospitals. However, important information can also be
obtained by 1-D scanning, with much less expensive spectrometers, using small
core samples taken in vivo at various processing stages.
268 Nuclear Magnetic Resonance
Table 6.9. Proton longitudinal relaxation time 'HT,p (ms) for solid pine wood, for PMDI/pine
and for UF/pine expressed as a function of the chemical shift (ppm) for lignin, hemicelluloses
and cellulose. (Marcinko et al. 1998, courtesy of Forest Products J)
7.00 r-----------------,
6.30
I0..
......
5.60
<> ?i...
........... '0
:t 4.90 .<5 .................... .
4.20 .~......
C"!
--
. ~ ••.... ::III.IEi·····I •••••••• •••i!!!!.
~
••.
3.50 L---L-----'-_.l...---'-----'_-'----'-_L---L-----'-----'
o 2 3 4 5 6 7 8 9 10 11
cure time (minutes)
Fig.6.41. The relaxation time 'H T,p (ms) for flakes cured with phenol formaldehyde resin at
110°,135° and 175°. (Schmidt and Frazier 2000; courtesy of Wood Fiber Science)
The challenge of NMR laboratory testing is to the analyze the adhesive bond
lines on pilot and industrial levels. Schmidt and Frazier (2000) demonstrated
the viability of cross-polarization magic angle spinning (CP/MAS) NMR mea-
surements for evaluating the degree of adhesive cure (phenol-formaldehyde)
in poplar flakeboards. Figure 6.41 shows the relaxation time IHT 1p (ms) for
flake samples cured with 13C phenol-formaldehyde resin. The curing tempera-
tures investigated were 110, 135 and 175°C. An increase in the relaxation time
1HT Ip with curing time and temperature was observed. These results give infor-
mation on the effects of temperature on the adhesive bond lines. CP/MAS
spectra of 13C flake samples of the skin and core of the boards at 13 and 24%
moisture contents are shown in Fig. 6.42. Skin samples are different from core
samples in the range of 35 and 63 ppm, which correspond to methylene and
hydroxymethyl, respectively. Skin samples have an increased methylene peak
and a diminished hydroxymethyl peak compared with core samples. The mean
ratios of corrected peak integrals of methylene carbon to the hydroxymethyl
carbon at 13% moisture content was for the core 0.8 and for skin, 1.6. At 24%
moisture content, these ratios were 1.0 and 1.8. Data for mean ratios of cor-
rected peak integrals for the core and skin were significantly different at K <
0.001. This difference is explained by the temperature difference between
the high temperature surface regions (200°C) and the cold face and the core
(150°C).
In conclusion, it can be noted that for pilot-scale flakeboards, the relative
degrees of resin cure are measurable by proton longitudinal relaxation time
eHT 1p ) and by the ratios of corrected peak resin integrals on CP/MAS spectra
of 13C nuclei.
270 Nuclear Magnetic Resonance
110 100 90 80 70 60 50 40 30 20 10
ppm
Fig. 6.42. CP/MAS spectra of BC-labeled flake samples of the core and of the face of the board at
13 and 24% moisture contents. (Schmidt and Frazier 2000; courtesy of Wood Fiber Science)
a) b) c)
Fig.6.43a-c. Kinetics of drying a water-based polyvinyl acetate adhesive emulsion between two
pine specimens as reported by Niemienen and Koenig (1988) after a lOmin; b 4h; c 16h. (VSP
International Science Publisher, with permission)
of time shows that after 3 min (Fig. 6.43a) a very thin glue line between the
wood specimens had been formed. The ll-mm trough in the specimen is also
evident on the right side of the wood sample.
After 4 h of drying (Fig. 6.43b), the adhesive line between the specimens is
dry and no longer visible on the slice. Only the still wet adhesive in the trough
is imaged. After 16h, the adhesive is still not completely dry, and an empty
space in the trough is visible, probably because the external pressure applied
to the adhesive joint was not sufficient. As expected, the adhesive strength of
the structure is decreased because of the void presence. The NMR signals are
due to water in the adhesive and not water that had penetrated from the adhe-
sive to the wood. The authors noted also that the mobility of thermoplastic
polymers increases at elevated temperatures and that NMR images can be
recorded even at 180 ac.
In routine industrial situations, voids in the adhesive layer must be detected
as early as possible to guarantee the quality of the adhesive bonds in mass pro-
duction offurniture, veneer, glue-lam, etc. The limitation of the proposed tech-
nique is related to the difficulty of detecting solid inclusions in the joints, such
as hard resins, because their very short relaxation times are not imaged. The
main advantage of NMR imaging for adhesive bonded structures is related
to the ease with which the images can be interpreted and to the informa-
tion provided about the adhesive itself and about the whole structure under
inspection.
272 Nuclear Magnetic Resonance
Another attractive idea is to use NMR imaging for the visualization of the
impregnation process of wood with preservatives or other liquids, that is, the
process of fluid absorption.
First, we will discuss the state of the art of the subject related to other
materials, such as fiber-reinforced composites and rocks, and then consider
measurements on wood.
Rothwell and Holecek (1984) demonstrated the feasibility of using NMR
imaging to follow the dynamics of water diffusion in inorganic materials. The
mechanism of water absorption depends on the physical and chemical para-
meters of the composite studied (glass fiber-reinforced epoxy-resin compos-
ite). Rapid characterization of the distribution of absorbed fluid in a polymeric
material is possible.
Kleinberg (1999) and many other authors before him, referred to the earlier
contributions that NMR had made to studies of the mesoscopic properties of
rocks, such as pore sizes and their distribution. Relaxation time measurements
can be used to determine a volumetrically weighted distribution of pore sizes
spanning several orders of magnitude. The quantitative measurement of the
amplitude of the NMR signal and the calibration of proton spin density mea-
surements provide a very accurate method to determine the porosity with a
noncontact and noninvasive technique. Similar methods can be used to inves-
tigate the porosity and penetrability of wood during preservation.
Hall and Rajanayagam (1986), using three-dimensional proton NMR
volume imaging, studied the impregnation of cylindrical maple specimens
of 1.2 cm diameter with preservative containing oxides of Cr, Cu and As and
with 1 M manganese chloride solution (MnCl z). The specimens were soaked
overnight with these substances. With a very coarse spatial resolution, Fig. 6.44
shows the relative intensity of the NMR signals from the "spin warp" technique,
measured on a sample compared with the spin lattice relaxation time T, of the
water protons. These images depend on the relationship between the pulse
intervals chosen for the image measurements and the characteristic relaxation
time T, of the water present in the wood. If the time between pulse sequences
is too short, the water protons in wood have not enough time to relax during
the measurement, and the image can have a low intensity. Moreover, a part of
the signal can disappear before it can be measured, if the signal from the water
has a short spin-spin relaxation time Tz The manganese chloride solution has
Tz = 10-5 s, much shorter than that of the water in the wood that is Tz = 10-3 s.
The Mn ions are magnetic and significantly reduce the T, and Tz relaxation
times of the absorbed fluids, which is an important technical aspect of these
measurements. Consequently, with a relatively short time between pulse
sequences, only a signal from the manganese chloride solution is observed.
This enables the diffusion of the MgCl z into the bath to be measured. The influ-
ence of impregnation time is shown in Fig. 6.45, where one can see the effect
after 3.5 and 0.5 h compared with the control measurements on a fresh wood
specimen.
Applications 273
a)
aDla
b)
c) fresh wood
d) impregnated wood
Fig.6.44a-d. Imaging with the "spin warp" technique of maple specimens impregnated with dif-
ferent preservatives, compared with a control specimen of fresh-cut wood, at 80MHz frequency
for protons and 15 min of total experimental time. The contour levels are logarithmically spaced.
The top contour level is at 0.7 of the tallest peak. The following level is at 0.7 of the previous level
height. The relative intensity of the NMR signal has been compared with the spin-lattice relax-
ation time of water protons. a Planar image of freshly cut maple. b Effect of heating one of them
for 3h at 1l0°C. c Planar image of three pieces of freshly cut wood. The piece in the middle has
been used as reference. d Piece soaked in a preservative that shows a substantial change. (Hall
and Rajanayagam 1986; courtesy of Wood Science Technology)
a)
Fig.6.45a,b. The influence of impregnation time with a 1-M solution of MnCI 2 on the distribu-
tion of the preservative in wood. a Fresh wood. b Sample after it had been soaked 3.5 and 0.5 h.
(Hall and Rajanayagam 1986; courtesy of Wood Science Technology)
Fig. 6.46. The CP/MAS spectrum of wood. (Lambert et a1.1991, with permission)
Fig. 6.46 with Fig. 6.2. Carbohydrate resonances are found with chemical shifts,
0, between 50 and 90 ppm. The lignin peaks are located between 0 = 100 ppm
and 0 = 160ppm. Archeological wooden artifacts, like elements of old ship-
wrecks, are of particular interest for the development of NMR imaging. These
objects survive only because they are water-logged and proton NMR imaging
is the most appropriate technique for studying their internal structure. The
very long-term action of water changes the physical and chemical properties
Applications 275
A 622-10 3s
• 437-10 3s
[J 292-10 3s
• 184-10 3s
V 107-10 3 s
... 55-10 3s
o 23-10 3s
• 7-10 3s
Fig. 6.47. One-dimensional NMR profile of the ingress of D20 in archeological oak (from the
Mary Rose ship) in longitudinal direction as a function of the depth of the sample_ (Robertson
and Packer 1999, with permission)
Fig. 6.48. The higher diffusion coefficient, as expected, was measured in the lon-
gitudinal direction. In this direction the diffusion pathway is along the fibers
and vessels and is similar to the pathways of fresh wood, but in an archeologi-
cal sample the diffusion can be blocked by deposits of calcareous material,
which can also block the diffusion through the medullary rays. In addition, the
permeability in the radial direction can be enhanced by the deterioration of pit
membranes. The permeability of archeological wood is considerably modified
by the degradation process. The solid core is less degraded and has a high
content of cellulosic components, but it is surrounded by a heavily degraded
surface layer. The variation of the diffusion coefficients in all anisotropic direc-
tions, expressed as a function of the depth of the specimen corresponds to a
typical pattern of deterioration of buried archaeological wood, with more
degradation at the surface than in the internal part of the specimen.
The potential areas of use to NMR imaging and spectroscopic techniques
are related to the identification of the tree species of archeological objects, esti-
mation of the age of samples from the ratio of unsaturated (lignin) to satu-
rated (carbohydrate) resonances, to assessment of the level of decomposition
of archeological wood prior to conservation and to the measurement of the
ingress of preservatives into the wood in order to stabilize it
NMR imaging for the measurement of the elastic properties of bulk solids was
first proposed and realized by Lewa (1991, 1992, 1996). This technique uses the
Applications 277
+· .
2.5 I
• longitudinal
. .
2.0
.. • tangential -
·
'·1- ... radial
~
~ 1·.-1
N
.L'.
51.5
.
'''''-
'"0 :, f·.-1
~
ci
1.0
,
.. ..... ~.!..j..... J-!-{ ................ f-•• j
.. f-H.
"~~~
0.5 ~~.~
·····················~·················t!=1
.......................
•
o I I 1
o 5 10 15 20 25 30
Oepht (mm)
Fig. 6.48. Diffusion coefficient of D,O as a function of depth and of anisotropic direction in
archeological oak from the Mary Rose ship. (Robertson and Parker 1999, with permission)
magnetic field gradient located in time and space and relates the spatial coor-
dinates with frequency. Such transformations are possible when the position
of an element or the value of the magnetic field intensity are time-dependent.
For a time-independent magnetic field with a constant gradient, the Larmor
frequency can be modulated by varying the posi,tion of the sample. This can
be accomplished by using a longitudinal mechanical wave. The modulation of
the magnetic field intensity affects the nuclei present in the displacement
element and depends on the argument of Bessel's functions. This argument is
expressed as a function of the angle between the displacement and the field
gradient vectors and the coefficient of the adiabatic compressibility of the
medium. The latter depends on the elastic properties of the solid and on instru-
mental factors such as: the type of paramagnetic nuclei, the adiabatic com-
pressibility and the density of the solid, and wavelength and velocity of the
mechanical wave.
The presence of the mechanical wave in the MRI field enables the display of
the bulk object by modulation of the Larmor frequency with a mechanical
wave applied to the sample in the presence of the magnetic field gradient. The
possibility of generating mechanical waves with wave-lengths comparable to
the MRI resolution and the cellular size, enables observations at a cell level.
Until today, the main application of this elastomagnetic resonance spec-
troscopy has been in the biomedical field, but extensions of such a technique
278 Nuclear Magnetic Resonance
6.5 Summary
NMR imaging is one of the most powerful and versatile techniques for the
characterization of materials. The development of high speed computers and
the introduction of high field superconducting magnets have increased the
speed, sensitivity and flexibility of such characterizations. Only the cost of such
equipment limits the utilization of NMR imaging in the fields of wood science
and technology. Standard NMR parameters are the resonance frequency,
the magnitude of the signal proportional to the density of the nuclei, the
spin-lattice relaxation time T 1, the spin-spin relaxation time T2 , the diffusion
coefficient, the flow velocity, the spin-spin coupling and the spectrum of cor-
relation times. The values of these parameters depend on species, moisture
content, physiological parameters of wood and on several instrumental and
measurement factors such as the Larmor precession frequency, temperature,
etc.
The NMR tomographic image depends on the scanning technique used,
pulse sequence, and magnetic field intensity. The NMR imaging technique
relies on the interaction of nuclear magnetic moment (nuclei) in only a small,
controlled zone of the sample under inspection, and is achieved by placing the
measured body in a spatially inhomogeneous field. Its nuclear resonance fre-
quency is matched to the RF signal only in the corresponding zone of the
object.
The NMR imaging technique can be used to investigate the spatial distri-
bution of all parameters that can be determined by NMR, such as densities, T 1>
T2 , diffusion terms, etc. Usually, the nuclear spin density and relaxation time
are mapped as a function of their spatial position. Using basic spatial encod-
ing and slice selection principles, different techniques are available to form
I-D, 2-D or 3-D images, using various spin-echo, stimulated-echo and
gradient-echo pulse sequences. NMR signals inherently depend on nuclear
relaxation time constant, which, in turn reflect the structural environment of
the emitting nuclei. There are several modalities of spatially encoding the
signals. One of these is to apply a linear magnetic field to the original static
field. In this way, nuclei on one side of the sample will feel a weaker total mag-
netic field than those on the other side. From such a set of data, the image of
the sample is reconstructed with an appropriate algorithm. Conventional NMR
spectroscopy can be coupled with the corresponding imaging technique, and
chemical structures of the specimen can be determined.
The main advantages of the NMR imaging technique are: the method is non-
destructive, non contact, relatively rapid, can be used in situ and in vivo, and
does not induce any structural damage. In addition to providing a relative
Summary 279
7.1 Introduction
loses, lignin, resins, etc.) can be neglected when this amount is compared with
the H present in the moisture of wood (Nakanishi et aI. 1998a).
Development of neutron imaging with computer tomography, in progress
in different laboratories (Furukawa et al. 2001), will give new opportunities in
the future to studies in wood science and tree physiology.
7.3 Equipment
Fig.7.1. Thermal neutron beam line (SINQ) at the Paul Scherrer Institute in Villigen, Switzer-
land. (Bauer 1998, with permission)
Fig. 7.2. Sample detector setup for neutron radiography. (Niemz et al. 2000c, with permission)
Applications 285
7.4 Applications
In this section, the imaging of the moisture in wood will be studied first on
disks taken from living trees during drying and then on small, clear specimens
during absorption of water For the study of drying phenomena, two cases will
be analyzed, corresponding to short and relatively long drying times. Imaging
of water distribution in structural elements submitted to weather condition-
ing will also be presented.
The storage of water in a tree is influenced by different factors, such as: the
capillarity of the wood structure and the elasticity of the cell tissue. The water
content in trees is subjected to seasonal changes and is different in sapwood
and in heartwood. Development of new methods for determination of the
water distribution in trees should be emphasized. Among the nondestructive
methods, neutron imaging seems to be promising. Water distribution was
studied with neutron radiography by Nakanishi and Watanabe (1995) who
used trees of different species and by Nakanishi et al. (1998b) with trees belong-
ing to the same species.
Table 7.2. Characteristics of the specimens irradiated for 2.3min. ( Data from Nakanishi and
Watanabe 1995, with permission)
d
Fig.7.3a-e. Photographic images and neutron radiography of different species. a Pinus
thunbergii. b Metasequoia glyptostroboides. c Chamecyparis obtusa. d Quercus serata. e Robinia
pseudoacacia. (Nakanishi and Watanabe 1995, with permission)
flux was obtained with (n, y) reaction and was 8.7 x 107 n/cm 2• The water dis-
tribution image was obtained on a film. The resolution of the image was
15 11m. The degree of brightness in the image was proportional to the water
content of the sample.
Having these images as a background, it can be noted that in Pinus thun-
bergii within the same ring, a heterogeneous water distribution has been
observed. The black zone in the pith indicates the decrease of water content,
suggesting the beginning of the necrosis of the tissue. In the image for Metase-
quoia glyptostroboides, we can see that only a small amount of water is
Applications 287
Nakanishi et al. (1998b) studied the moisture distribution in the inner part of
the sapwood of different cultivars of sugi (Cryptomeria japonica) with special
emphasis on the identification of the presence of a so called white zone in the
sapwood, characterized by a high moisture content. This zone is probably
caused by genetic factors and environmental conditions. It is present in all
specimens, as shown in Fig. 7.4.
In three-dimensional representation of water distribution, the aspect of the
"white zone" is very well defined for all specimens (Fig. 7.5). In addition, near
the bark, this distribution of moisture is not superimposed on the pattern of
the annual rings. Nakanishi and Matsubayashi (1997) suggested that this bright
zone observed by neutron radiography could also indicate the presence of dif-
ferent chemical components rich in H, which are different from those in water
from the cell wall.
In conclusion, it can be said that with neutron radiography, the visualiza-
tion of a patterns of water distribution across the transverse section of a tree,
mainly near the bark, was possible for all species studied, Neutron imaging
allows studies on water distribution in zones near the bark, which are difficult
to perform with other techniques.
288 Neutron Imaging
Fig. 7.4. Photographic images (A, B, C, D) and corresponding neutron images (a, b, c, d) of dif-
ferent cultivars of sugi (Cryptomeria japonica). A 24-year-old cultivar, 25 Gou. B 25-year-old cul-
tivar Honjiro. C 29-year-old cultivar I-Gou. D 3D-year-old cultivar, Sanbusugi. (Nakanishi et al.
I998b, with permission)
Established wood drying techniques are based on the removal of the large
amount of moisture present in green wood and have the purpose of protect-
ing wood materials from excessive drying stresses that can cause defects and
degradation. For wood quality control during drying, several nondestructive
methods have been developed. In this section, we will discuss some laboratory
measurements undertaken with the neutron technique for imaging the kinet-
ics of drying.
Applications 289
Fig. 7.5. Three-dimensional water distribution in a disk of l-cm thickness of green Cryptomeria
japonica. (Nakanishi and Matsubayashi 1997, with permission)
'"a
~
a
8::l
S
~
&'
a)
Fig.7.6a. Imaging of the kinetics of moisture content distribution. a For 6h drying at 60°C at 90% relative humidity of four disks (A, B,
C, D as in Fig. 7.4) of I-cm thickness of Cryptomeria japonica.
Applications 291
255
-
c:
;:,
0
0
255
E
I'll
GI 0
>
:;: 255
I'll
...
li
0
255
0
b) s S H S S H S
Fig.7.6b. Variation of the moisture content profile of four disks A, B, C, D for which the numbers
0,2,4,6 are the drying hours. (Nakanishi et al. 1998b, with permission)
Fig. 7.7. Neutron radiography of spruce drying as a function of time. (Lehmann et al. 2001, with
permission)
A, but after 6 h of drying, the moisture content is higher in the sapwood than
in the heartwood. For all these three disks the heartwood represents about 33%
of the transverse section. By contrast, for disk D, the zone occupied by the
heartwood seems to amount to about 70% of the transverse section. After 6 h
of drying, the heartwood lost a considerable amount of water compared with
the sapwood. The causes of the different moisture distributions are unknown,
but the effects of such differences on the wood during drying can be disastrous
because of the development of a nonuniform drying stress.
A longer drying process (45 h) was studied with spruce specimens (Fig. 7.7) by
Lehmann et al. (2001). During the drying process at 20°C and 65% relative
292 Neutron Imaging
humidity, the wood moisture content decreased from a saturated state to air
dry at about 18%. After 22 h of drying, some anatomic details at the transverse
section of specimens were observed. After 4S h, the disposition of the annual
rings was clearly visible, together with shrinkage of the specimen. The regular
square cross section of the specimen in the water-saturated state had been
transformed during drying into a rhomboid-shaped section. The minimum
shrinkage was along the radial direction of the wood (at 4S0 in the figure)
which corresponds to the direction of the rays. The maximum shrinkage
occurred in the tangential direction to the annual rings. The drying kinetics
(Fig. 7.8) have been investigated by advanced image processing, by subtract-
ing image characteristics in two different drying stages. The image of the sat-
urated sample has been considered as a reference. During drying, a crack
observed in the upper part of the image (after 8h of drying) increased in size
and propagated in the tangential direction to the annual ring. After 70 h of
drying, the crack is well visible on the external, dried area of the specimen.
Wood can be protected from the attack of biological agents by applying chem-
ical preservatives. The uniform distribution and sufficient penetration of
these substances can greatly increase the life of wood structures. For an under-
standing of the complex phenomena during the impregnation and for the
determination of preservative effectiveness, it is necessary to utilize nonde-
structive techniques able to visualize in real time the absorption of liquids by
the materials. This section deals with the determination of water flow in solid
wood and wood-based composites with the neutron imaging technique.
Different wood species (beech, pine, spruce, sweet chestnut), particleboard
and MDF specimens (Lehmann et al. 2001) were partially immersed in water,
as shown in Fig. 7.9. Quantitative water absorption as a function of time is
shown in Fig. 7.10. After Sh of immersion, the beech samples absorbed 199
of water, while all other specimens absorbed about 4 g of water. After 22 h of
immersion, the maximum water absorption was 2S g for beech. During the
same period, water absorption was at a minimum, (about 1 g) for the MDF
Applications 293
I mill walcr
Fig. 7.9. Experimental device for water absorption by capillarity in wood samples 200 mm long
x lOmm wide x 16mm thick. (Lehmann et al. 2001, with permission)
30
_ _ beech
25
~
/
/
5
/ _partie leboard
_ spruce
o
IL ~
pine
MDF
o 5 10 15 20 25
time (hours)
Fig. 7.10. Water absorption as a function of time for beech, spruce, pine, chestnut, particleboard
and MDF samples. (Lehmann et al. 2001, with permission)
sample. The kinetics of water absorption for beech, pine and particleboard
samples, with neutron radiography as a function of time, is shown in Fig. 7.11.
On the image, the darker zone, which corresponds to the maximum amount of
absorbed water, is observed for the beech samples.
a)
b)
cult to match with other materials, when architectural considerations are taken
into account. The importance of a proper connection design cannot be over-
stated (Natterer et al. 2000).
On the connections, the effect of weathering may be undesirable, depend-
ing on the requirements for a particular wood product. Natural weathering can
produce on wood connections a nonuniform moisture content distribution
that induces shrinkage and swelling. The accompanying stress thus induced by
shrinkage and swelling causes excessive dimensional changes, such as warping,
surface roughening and checking, loss of some surface fibers, etc. The physi-
Applications 295
Table 7.3. Specimen characteristics for neutron radiography. (Data from Niemz et al. 2000c, with
permission)
Dry
Wet
.)
Dry
Fig.7.12a,b. Moisture content distribution in corner connections for dry and moisturized
samples. a Spruce/spruce doweled. b Obtuse insertion spruce/beech, glued. (Niemz et al. 2000c,
with permission)
Summary 297
0.25
~ joint joint
.r: 0.2
c..
0)
"0 _ dry specimen
"0 0. 15 wet specimen
0)
.~
<a 0.1
E
0
spruce bccch sprLIce
c 0.05
"1 Of' <'! r--: <'! r--: <'! r-:
.",. r'l <'l N r'l
Of,
N N
r-:
'"
N <'l .",. '0 00 0
a) distance (mm)
M 0.16
E
<) 0.14 I
Ob
'-' 0.12 I
c 0.1 I
.B 0.08 1
13
u 0.06
~ 0.04
::l
'iii 0.02 1
'0
E 0
0 20 40 60 80 100 120
b) distance (mm)
Fig. 7.13. Moisture distribution profile as a function of the geometry of the joint for obtuse inser-
tion (spruce/beech). a Normalized depth as a function of distance (length of the specimen).
b The corresponding moisture content as a function of distance. (Niemz et al. 2000c, with
permission)
7.5 Summary
Neutron beam imaging of wood structure is one of the most recent nonde-
structive techniques developed for this material. The neutrons are subatomic
particles with zero charge. They are the major constituents of nuclei and are
produced in fission reactors. Neutrons in thermal equilibrium with a moder-
ator material near room temperature are called thermal neutrons. Neutron
radiography of wood provides a specific hydrogen image of high resolution
(IS ~m) and allows the imaging of water distribution in wood and in living
wood cells. The interaction of neutrons with wood is weak, and the absorption
is small and, consequently, large samples can be inspected. Today, the applica-
298 Neutron Imaging
Bauer G (1998) Operation and development of the new spallation neutron source SINQ at Paul
Scherrer Institut. Nuclear Instrum MethodsPhys Res B 139:65-71
Bauer G, Dai Y, Wagner W (2001) SINQ, Layout, operation applications and R&D to high power.
Proc Symposium Structural materials for hybrid systems: a challenge in metallurgy, Paris,
29-31 October,6 pp
Bauer M, Guntrum C, Ota M, Rippel W, Busse G (1992) Thermographic characterization of defects
and failure in polymer composites. In: Balageas D, Busse G, Carlomagno GM (eds) Proc Quan-
titative Infrared Thermography (QIRT92). Editions Europeennes Thermiques et Industrie,
Paris, pp 141-144
Beall F (1996) Future of nondestructive evaluation of wood and wood-based materials. Holz-
forsch Holzverwert 5:73-75
Beatty R, Choffel D, Charpentier P (1999) Towards products tracking into wood industries. Proc
IWMS:325-334
Beaudoin JL, Merienne R, Danjoux R, Egee M (1985) Numerical system for infrared scanners and
application to the subsurface control of materials by photothermal radiometry. Infrared Techn
Appl SPIE 590:287
Bennet CA, Patty RR (1982) Thermal wave interferometry: a potential application of the pho-
toacoustic effect. Appl Opt 21:49-54
Benson-Cooper DM, Knowles RL, Thomson FJ, Cown DJ (1982) Computed tomographic scan-
ning for the detection of defects within logs. FRI Bull For Res Inst N Z 8:9
Beravs K, Oven P, Sersa I, Torelli N, Demsar F (1998) Electric current density imaging of pedun-
culate oak (Quercus robur 1.) twigs by magnetic resonance imaging. Holzforschung 52:
541-545
Beres C, Fenyvesi A (1994) The role of computer tomographical methods in investigation of water
transport in oak trees. Proc 1st Eur Symp on Nondestructive Evaluation of Wood. University
of Sopron, Sopron, Hungary, 525
Berglind H, Dillenz A (2000) Detection of glue deficiency in laminated wood with thermography.
12th Symp Nondestructive Testing of Wood, University of Western Hungary, Sopron, pp 413-
420
Bernard P, Chalebois A, Pearson N, Samson M (1996) Application of infrared ellipsometry to
detection oflocal grain organization and spiral grain in wood. 10th Symp NDT of Wood. Press
Univ Romandes, Lausanne, pp 165-174
Bernatowicz G (1994) Determination of moisture distribution and check formation by means of
thermography. (Bestimmung des Feucheverteilung und Rissbildung mittels Thermographie).
Ann Warsaw Agric Univ For Wood Technol 45:65-69
Berndt H, Johnson GC (1994) Examination of wave propagation in wood from a microstructural
perspective. In: Thompson DO, Chimenti DE (eds) Review of progress in quantitative nonde-
structive evaluation, vol 14b. Plenum Press, New York, pp 1661-1668
Berndt H, Schniewind AP, Johnson GC (1999) High resolution ultrasonic imaging of wood. Wood
Sci TechnoI33:185-198
Berndt H, Schniewind AP, Johnson GC (2000) Ultrasonic energy propagation through wood:
where, when how much. Proc 12th Symp Nondestructive Testing of Wood. University of
Western Hungary, Sopron, pp 57-65
Bhardwaj MC (2000) High transduction piezoelectric transducers and introduction of noncon-
tact analysis. In: Harvey JA (ed) Encyclopedia of smart materials. Wiley, New York, 25 pp
Bhardwaj MC, Neeson I, Langron ME, Vandervalk E (2000) Contact free ultrasound: the final fron-
tier in nondestructive materials characterization. Paper at 24th Conference: an International
conf on engineering ceramics and structures. 24-27 Jan 2000, The American Ceramic Society,
Cocoa Beach, FL, USA
Biagi E, Gatteschi G, Masotti L, Zanini A, Cerofolini M, Lorenzi A (1994) Tomografia ad ultrasuoni
per la caratterizzazione difettologica dellegno. (Ultrasonic tomography for defect character-
ization in wood). Alta Frequ Riv Elettronica 6 2:48-57
Biernacki JM, Beall FC (1993) Development of an acousto-ultrasonic scanning system for non-
destructive evaluation of wood and wood laminates. Wood Fiber Sci 25:289-297
References 301
Birkeland R (1985) The status of tomographic scanning as a tool for detecting internal log defects.
Proc 5th Nondestructive Testing of Wood Symposium, Washington State University. Eng Pub-
lications, Pullman, pp 231-237
Birkeland R (1990) NDE facilities and research program in Norway. 7th Symp on Nondestructive
Testing of Wood, Washington State University, Madison, pp 57-62
Birsk AS (1972) Particleboard blow detector. For Prod J 22 6:23-26
Bodig J (1994) NDE of wood in North America. Concept and applications. Proc First Eur
Symp on Nondestructive Evaluation of Wood. University of Sopron, Sopron, Hungary, pp
1-l3
Bodig J, Jayne BA (1982) Mechanics of wood and wood composites. Van Nostrand Reinhold,
New York, 712 pp
Bolomey JC, Pichot C (1990) Microwave tomography: from theory to practical imaging systems.
Int J Imag Syst TechnoI2:144-156
Bourbie T, Coussy 0, Zinszner B (1987) Acoustics of porous media. Gulf Publ Co, Houston, TX,
334 pp
Bowyer JL (2000) Wood science in a changing world. Where are we headed? Wood Sci Technol
34:175-181
Brown DP, Pratum TK, Bledsoe C, Ford ED, Cothern JS, Perry D (1991) Nonivasive studies of
conifer roots: nuclear magnetic resonance imaging of Douglas fir seedlings. Can J For Res 21:
1559-1566
Brown JH, Davidson RW, Skaar C (1963) Mechanism of electrical conduction of wood. For Prod
J Oct:455-459
Brownstein KR (1980) Diffusion as an explanation of observed NMR behavior of water absorbed
on wood. J Magn Reson 40:505-510
Buckley J (2000) Air-coupled ultrasound:... a millennial review. 15th World Conf On NDT, Roma
http://www.ndt.netlarticle/wcndtOO/paper/idn507/idn507.htm
Bucur V (1980) Anatomical structure and some acoustical properties of resonance wood. Catgut
Acoust Soc Newsl, series I, no 33:24-29
Bucur V (1986) Les termes non diagonaux de la matrice des rigidites du bois. (Off-diagonal terms
of stiffness matrix of wood) Holzforschung 40:315-324
Bucur V (1990) NMR imaging of green wood. INRA Document a distribution limite, Institut
National de la Recherche Agronomique, Champenoux, France, 10 pp
Bucur V (1992) Anisotropy characterization of structural fiakeboards with ultrasonic methods.
Wood Fiber Sci 24:337-346
Bucur V (1995) Acoustics of wood. CRe Press, Boca Raton, 284 PP
Bucur V, Archer RR (1984) Elastic constants of wood by an ultrasonic method. Wood Sci Technol
18:255-265
Bucur V, Rasolofosaon PNJ (1998) Dynamic elastic anisotropy and nonlinearity in wood and rock.
Ultrasonics 36:8l3-824
Bucur V, Rocaboy F (1987) Surface wave propagation in wood: prospective method for the deter-
mination of wood off-diagonal terms of stiffness matrix. Ultrasonics 26:344-347
Bucur V, Lanceleur P, Roge B (2001) Acoustic properties of wood in tridimensional representa-
tion of slowness surfaces. Ultrasonic International Symp Delft, p 6
Burov VA, Rumyantseva OD (1996) Linearized inverse problem of scattering in monochromatic
and pulse modes. Acoust Phys 40:34-42
Busse G (1979) Optoacoustic phase angle measurement for probing a metal. Appl Phys Lett 35:
759-760
Busse G (1988) Imaging with optically generated thermal waves. Physical acoustics, vol XXII.
Academic Press, New York, pp 403-478
Busse G (1994) Nondestructive evaluation of polymer materials. NDT Eval Int 27 5:253-
262
Busse G (2001) Lockin thermography. In: Maldague X (ed) Nondestructive testing handbook, vol
3. Infrared and thermal testing. American Society for Nondestructive Testing Inc, Columbus,
OH, pp 318-327
302 References
Chang SI, Cohen M, Wang PC (1989a) Ultrafast scanning of hardwood logs with a NMR scanner.
Proc 3rd Int Conf On Scanning Tech, Sawmilling
Chang SI, Olson IR, Wang PC (1989b) NMR imaging of internal features in wood. For Prod I 39
6:43-49
Charles lA, Wilson DW (1981) A model of passive thermal nondestructive evaluation of com-
posite laminates. Polym Composites 2 3:105-111
Chazelas IL (1991) Caracteristiques physiques et mecaniques locales du bois dans la zone des
noeuds. These de Doctorat Universite Blaise Pascal, Clermont Ferrand, 156 pp
Chazelas IL, Vergne A, Bucur V (1988) Analyse de la variation des proprietes physiques et
mecaniques locales du bois autour des nreuds. (Wood local properties variation around
knots). Actes du Colloque "Comportement Mecanique du Bois". GS Rheologie du Bois,
Bordeaux, France, pp 344-347
Chelidze T, Gueguen Y, LeRavalec M (1998) From classic fractal mechanics of disordered media:
self-consistency versus self-similarity. In: Frantziskonis GN (ed) Proc PROBAMAT - 21st
century: probability and materials. Test, models and applications for the 21st century. Kluwer,
Dordrecht, pp 197-231
Chiorescu S, Gronlund A (2000) Validation of a CT-based simulator against a sawmill yield. For
Prod I 50 6:69-76
Choffel D (1999) Automation of wood mechanical grading. Coupling of vision and microwave
devices. SPIE 3836:114-121
Choffel D, Martin P (1996) Microwaves and vision device for mechanical grading. Proc 10th Symp
Nondestructive Testing of Wood, Press Polytechniques et Universitaires Romandes,
Lausanne, pp 331-339
Choffel D, Goy B, Gapp D (1992) Interaction between wood and microwaves automatic grading
application. Workshop on scanning technology and image processing on wood. Lulea Uni-
versity of Technology, Lulea, Sweden, pp 1-8
Choffel D, Martin P, Gross P (1995) Nondestructutive testing methods on solid timber:
microwaves, ultrasound and spectrometry. Proc Int Conf Progress in Forest Products
Research, Gottingen, pp 86-89
Choong ET (1963) Movement of water through a softwood in the hygroscopic range. For Prod I
13:489-498
Cielo P, Krapez IC, Lamontagne M (1988) Lumber moisture evaluation by a reflective cavity pho-
tothermal technique. Rev Phys AppI23:1565-1576
Coates ER, Chang SI, Liao TW (1998) A quick defect detection algorithm for magnetic resonance
images of hardwood logs. For Prod J 4810:68-74
Cohen MH, Mendelson KS (1982) Nuclear magnetic resonance and the internal geometry of sed-
imentary rocks. I Appl Phys 53 2:1127-1135
Cole-Hamilton DI, Chudek lA, Hunter G, Martin CJM (1990) NMR imaging of water in wood
including water-logged archaeological artefacts. I Inst Wood Sci 12:111-113
Coles ME (1999) X-ray imaging. In: Wong P-Z (ed) Experimental methods in the physical sci-
ences, vol 35. Academic Press, San Diego, pp 301-336
Comino E, Socco V, Martinis R, Nicolotti G, Sambuelli L (2000) Ultrasonic tomography for wood
decay diagnosis. In: Backhauss GF, Bader H, Idezak E (eds) Int Symp Plant Health in Urban
Horticulture. Mitt. Bundesanst Land-Forstwirtschaft, Braunschweig, p 279
Cormack AM (1963) Representation of a function by its line integrals with some radiological
applications. Part I: I Appl Phys 34:2722-2727; Part II: I Appl Phys 35:2908-2913
Corral CZ, Lindquist CS (1998) On implementing Kasa's circle fit procedure. IEEE Trans Instrum
Meas 47:789-795
Cown DI, Clement BC (1983) A wood densitometer using direct scanning with X-rays. Wood Sci
TechnoI17:91-99
Crawford CR, Kak AC (1982) Multipath artifacts in ultrasonic transmission tomography. Ultra-
sonic Imaging 4:234-266
Davis WC, Moschler Ir (1986) A direct scanning densitometer to measure density profiles in wood
composite products. For Prod J 36, 11/12:82-86
304 References
Davis JR, Weels P, Morgan M, Shadbolt P (1989) A field portable X-ray CT pole scanner and CT
log scanning. Proc of 7th Nondestructive Testing of Wood Symposium, Washington State Uni-
versity. Eng Publications, Pullman, pp 251-262
Davis JR, Lerdin A, Wells P, Ilic J (1991) X-ray microtomography of wood. J Int Wood Sci, 12,4:
259-261
Dean SR (1983) The Radon transform and some of its applications. Wiley, New York, 289 pp
Delsanto PP, Romano A, Scalerani M, Moldoveanu F (1998) Application of genetic algorithms to
ultrasonic tomography. J Acoust Soc Am 104:l374-1381
Dench EC (1973) Advantages of microwave processing. Trans Int Microwave Power Inst 1:l3-16
Denis F, Jossinet J, Gimenez G (1986) Tomographie acoustique: etats, applicatiations et per-
spectives. Rev Acoust 79:21-25
Devaney AJ (1982) A filtered back propagation algorithm for diffraction tomography. Ultrasonic
Imaging 4:336-350
Devaney AJ (1986) Reconstructive tomography with diffractive wave fields. Inv Prob 2:161-
183
Dillenz A, Wu D, Breitriick K, Busse G (2000a) Lock-in thermography for depth
resolved defect characterization. 15th World Congress on NDT, Roma, 6 pp,
http://www.ndt.net/article/wcndtOO/papers/idn3111idn311.htm
Dillenz A, Zweschper T, Busse G (2000b) Elastic wave burts thermography for NDE of subsurface
features. Insight 42 12:815-817
Dines KA, Lytle RJ (1979) Computerized geophysical tomography. Proc IEEE 67 7:1065-1073
Divos F, Szegedi S, Raics P (1994) Local densitometry of wood by gamma back-scattering. Holz
Roh Werkst 54:279-281
Domanus JC (1992) Practical neutron radiography. Kluwer, Boston
Douglas Mast T (1999) Wideband quantitative ultrasonic imaging by time-domain diffraction
tomography. J Acoust Soc Am 1066:3061-3071
Dreele von RB (1994) Neutron diffraction. In: Cahn RW, Hassen P, Kramer EJ (eds) Material sci-
ences and technology. A comprehensive treatment, vol 2B. VCH, Weinheim, pp 562-609
Drobakhin 0 (2000) Homographic approach to multifrequency microwave methods of dielectric
material NDE. 15th World Conf on Nondestructive Testing, Roma, 5 pp
Dubey YM, Deorani SC (1997) Dielectric properties of coniferous timbers at microwave fre-
quencies. J Indian Acad Wood Sci 1:77-82
Dumoulin JP, de Belleval JF (1987) Flaw classification in composites using ultrasonic echography.
In: Alippi A, Mayer W (eds) Ultrasonic methods in evaluation of inhomogeneous materials.
Martinus Nijhoff, Dordrecht, pp 207-217
Elbez G (1990) Medical X-ray scanner applied to wood gluing. Centre Technique du Bois et de
l' Ameublement, Paris, 5 pp
EI-Rayes MA, Ulaby FT (1987) Microwave dielectric spectrum of vegetation. Part I: experimen-
tal observations. IEEE Trans Geosci Remote Sensing GE 25 5:541-549
Falls S, Chow T, Yong R, Hutchins D (1989) Acoustic emission analysis and ultrasonic velocity
imaging in the study of rock failure. J Acoust Emission 8:S166-S169
Faust TD, Tang M, Bhandarkar S, Smith JW, Toliner EW (1996) Evaluation of growth related
features in selected hardwoods logs using X-ray computed tomography and image analysis.
Proc lOth Symp Nondestructive Testing of Wood, Presses Poly techniques et Universitares
Romandes, Lausanne, pp 201-208
Fengel D (1991) Aging and fossilization of wood and its components. Wood Sci Technol 25:
153-177
Feraz ES de B (1976) Wood density determination by attenuation of low-energy gamma radia-
tion. IPEF Piracicaba Bresil 12:61-68
Filippini B, Lambert P, Martin P, Mallick G, Leban JM (1990) Research on wood properties and
defects with microwave probe (Investigation des proprielreIs du bois et des singularitels par
capteur micro-ondes). Actes du 3-e2me Colloque. Sciences et Industries du Bois. ARBORA,
Bordeaux Tome I, pp 1l3-121
Fink M (1983) Ultrasonic imaging (L'imagerie ultrasonore). Rev Phys AppI18:527-556
References 305
Fioravanti M, Ricci R (1991) L'impiego della tomografia computerizzata per misure densitomet-
riche sullegno: indagine sperimentale e risultati metodologici. (Use of computed tomogra-
phy for densitometric measures on wood: experimental researches and methodological
results). Ann Acad Ital Sci For XL:I-35
Flibotte S, Menon RS, MacKay AL, Hailey JRT (1990) Proton magnetic resonance of western red
cedar. Wood Fiber Sci 22:362-376
Forrer JB, Funck JW (1998) Dielectric properties of defects on wood surfaces. Holz Roh Werkst
56:25-29
Fortunko C, Renken M, Murray A (1991) Examination of objects made of wood. Proc IEEE Ultra-
sonic Symp, pp 1099-1103
Frantziskonis GN, Blodgett MP (1998) Multiscale material characterization and applications.
In: Frantziskonis GN (ed) Proc PROBAMAT - 21st century: probability and materials. Test,
models and applications for the 21st century. Kluwer, Dordrecht, pp 367-378
Frick J, Sawade G, Schuh H (1997) Defect localization in pillars of marble by ultrasonic tomog-
raphy. Otto GrafJ 8:174-188
Fridel MJ, Tweeton DR, Jackson Jessop JA, Billington S (1992) Mining application of seismic
tomography. 62nd Annu Int Meet Soc Exploration Geophys BG 1 7, pp 58-62
Fukada E (1965) Piezoelectric effect in wood and other crystalline polymers. 2nd Symp NDT of
wood, Washington State University, Pullmann, pp 143-170
Fukada E, Yasuda S, Kohara J, Okamoto H (1956) Dynamic Young's modulus and piezoelectric
constants of old timber. Bull Kabayasi Inst Phys Res 6:104-107
Funt BV, Bryant EC (1985) A computer vision system that analyses CT-scans of sawlogs. Proc of
IEEE Computer Soc Conf on Computer vision and pattern recognition, San Francisco, CA,
9-13 June,pp 175-177
Funt BV, Bryant EC (1987) Detection of internal log defects by automatic interpretation of com-
puter tomography images. For Prod J 37 1:56-62
Furukawa J, Nakanishi TM, Matsubayashi M (2001) Water uptake activity in soybean root revealed
by neutron beam imaging. Nondestr Test EvaI16:335-334
Gadian DG (1982) Nuclear magnetic resonance and its applications to living systems. Clarendon
Press, Oxford, 197 pp
Garrett DA (1977) The technological development of neutron radiography. Atomic Energy Rev
15:125-142
Garrett DA, Berger H (1977) The technological development of neutron radiography. Atomic
Energy Rev 15,2:125-142
Gierlik E, Dzbenski W (1996) Wood densitometry based on the radiometric methods. Proc 10th
Symp Nondestructive Testing of Wood, Presses Poly techniques et Universitaires Romandes,
Lausanne, pp 217-225
Gil AM, Neto CP (1999) Solid state NMR studies of wood and other lignocellulosic materials.
Annu Rep NMR Spectrosc 37:75-117
Gilardi G, Abis L, Cass AEG (1995) Carbon 13 CP/MAS solid-state NMR and FT-IR spectroscopy
of wood cell wall biodegradation. Enzyme Microbial TechnoI17:268-275
Gilbert P (1972) Iterative method for the tridimensional reconstruction of an object from pro-
jections. J Theor BioI 36: 105-117
Gilboy WB (1984) X- and y-ray tomography in NDE applications. Nuclear Instrum Meth Phys Res
221:193-200
Gilboy WB, Foster J (1982) Industrial applications of computerized tomography with X- and
gamma-radiation. In: Sharpe RS (ed) Research techniques in nondestructive testing, vol 6.
Academic Press, London, pp 255-287
Giordano G (1971) Tecnologia dellegno. Vol I. Materia prima. Unione Tipografico - Editrice
Torinese, Milano, 1086 pp
Glos P (1982) Machine grading of sawn timber: current technology and comparison of methods.
Holz ZentralbI108:153-155
Golosovsky M, Lann AF, Davidov D, Frenkel A (2000) Microwave near-field imaging of conduct-
ing objects of a simple geometric shape. Rev Sci Instrum 71 10:3927-3932
306 References
Goy B, Martin P, Leban JM (1992) The measurements of wood density by microwave sensor. Holz
Roh Werkst 50:163-166
Granier A, Anfodillo T, Sabatti M, Cochard H, Dreyer E, Tomasi M, Valentini R, Breda N (1994)
Axial and radial water flows in trunks of oak trees: a qualitative and quantitative analysis.
Tree Physiol 14: 1383-1396
Greaves H (1998) Trends in wood research and utilization - a view from the institute of wood
science, Australian branch. J Inst Wood Sci 146:293-300
Green RE Jr (1987) Ultrasonic nondestructive materials characterization. In: Vary A (ed)
Materials analysis by ultrasonics. Noyes Data Corp, Park Ridge, USA, pp 1-29
Greenleaf JF (1981) Computerized transmission tomography. In: Marton L, Marton C (eds)
Methods of experimental physics, vol 19. Ultrasonics. Academic Press, New York, pp 563-589
Grimberg R, Savin A, Lupu A, Rotundu C, Iancu L (2000) A method to determine the debonding
zones in multilayer wood materials. 15th World Conference on Nondestructive Testing, Roma,
2000,6 pp
Gronlund A, Grundberg S, Gronlund U (1994) The Swedish stem bank - an unique database for
different silvicultural and wood properties. IUFRO S5.01-04 Workshop Proc, Hook, Sweden,
pp 71-77
Grundberg S, Gronlund A (1997) Simulated grading of logs with an X-ray log scanner. Grading
accuracy compared with manual grading. Scand J For Res 12:70-76
Grundberg S, Gronlung A, Gronlund U (1995) The Swedish Stem Bank - a data base for different
silvicultural and wood properties. Res Rep TULEA [BR3] 1995:31. Lulea Unic Technol,
Skelleftea, Sweeden, 14 pp
Grundberg S, Lindgren 0, Gronlung A (1996) The development of an X-ray based scanner for
log grading at full production speed. Proc 10th Symp Nondestructive Testing of Wood Swiss
Federal Institute of Technology, Press Poly techniques et Universitaires Romandes, Lausanne,
p 407
Guinn VP, Lukens HR (1965) Nuclear methods. In: Morrison GH (ed) Trace analysis: physical
methods. Interscience, New York
Habermehl A( 1982a) A new nondestructive method for determining internal wood condition and
decay in living trees. I. Principles, method and apparatus. Arboric J 6:1-8
Habermehl A (1982b) A new nondestructive method for determining internal wood condition
and decay in living trees. II. Results and further developments. Arboric J 6:121-130
Habermehl A, Ridder HW (1992a) Metodik der Computer - Tomographie zur zerstOrungsfreien
Untersuchung des Holzkorpers von stehenden Biiumen. (Computerized tomography applied
to studies on wood of living trees). Holz Roh Werkst 50:465-474
Habermehl A, Ridder HW {1992b) Computer Tomographie am Baum (Computed tomograhy for
tree). Materialpriifung 34:325-329; 357-360
Habermehl A, Ridder HW (1993) Anwendungen der mobilen Computer-Tomographie zur zer-
stOrungsfreien Untersuchung des Holzkorpers von stehenden Biiumen (Application of mobile
computed tomography for nondestructive investigations on wood of living trees) Holz Roh
Werkst 51:1-6
Habermehl A, Ridder HW (1994) Applications of computerized tomography in forest and tree
sciences. Int Symp on Computerized Tomography for Industrial Applications, 8-10 June,
Berlin, pp 71-81
Habermehl A, Ridder HW (1995) Computerised tomographic investigations of street and park
trees. Arboric J 19:419-437
Habermehl A, Ridder HW (1996) Computer Tomographie in der Forstwirtschaft und Baumpflege.
(Teil1 und 2) DGZfP/DACH - Zeitung no 55:48-55 and no 56:47-55
Habermehl A, Ridder HW (1998) y-ray tomography in forest and tree sciences. Proc SPIE (lnt Soc
Optical Eng) 3149:234-244
Haegglund G, Lindgren 0 (1985) Maetning av densitets-och fuktvariationer i traematerial med
hjaelp av datortomografi. (Measuring of density and moisture variations in wood with the
help of computed tomography). Traeteknikrapport no 67. Stockholm Traeteknik Centrum,
Stockholm, 21 pp
References 307
Hagman POG, Grundberg SA (1995) Classification of scots pine (Pinus sylvestris) knots in density
images from CT scanned logs. Holz Roh Werkst 53:75-81
Hailey JRT, Menon RS, Mackay A, Burgess AE, Swanson JS (1985) Nuclear resonance scanning for
log characterization. 5th Symp Nondestructive Testing of Wood, Washington State University,
Pullman
Halabe UB, Ganga Rao H, Rao Hota V (1995) Nondestructive evaluation of wood using frequency
domain analysis. In: Thompson DO, Chimenti DE (eds) Review of progress in quantitative
nondestructive evaluation, vol 14. Plenum Press, New York, pp 1653-1660
Hall LD, Rajanayagam V (1986) Evaluation of the distribution of water in wood by use of three
dimensional proton NMR volume imaging. Wood Sci Technol 20:329-333
Hall LD, Rajanayagam V, Stewart WA, Steiner PR (1986a) Magnetic resonance imaging of wood.
Can J For Res 16:423-426
Hall LD, Rajanayagam V, Stewart WA, Steiner PR, Chow S (1986b) Detection of hidden morphol-
ogy of wood by magnetic resonance imaging. Can J For Res 16:684-687
Halloin JM, Cooper TG, Potchen EJ, Vozzo JA (1994) Magnetic resonance imaging, a technology
for noninvasive plant analysis. In: Research and applications of chemical sciences in forestry.
USDA, GTR SO-104, New Orleans, Louisiana, 9 pp
Harley J, Morris PI (1988) Application of scanning and imaging techniques to assess decay
and wood quality in logs and standing trees. Forintek Canada Corp no FO 42-91-42-(1988)E,
10 pp
Harpole GB, McDonald KA (1981) Investment opportunity: a scanning - ultrasonics cut stock
manufacturing system. USDA, FPL, Research Paper 390, 7 pp
Hatcher PG, Berger lA, Earl WL (1981) NMR studies of ancient buried wood. I. Observations on
the origin of coal to the brown coal stage. Org Geochem 3:49
Hattori Y, Kanagawa Y (1985) Nondestructive measurement of moisture distribution in wood with
a medical X-ray CT scanner. Part I. Mokuzai Gakkaishi 31:974-982
Hauffe P, Mahler G (2000) Evaluation of internal log quality using X-ray and ultrasound. 12th
Symp Nondestructive Testing of Wood, Sopron, pp 259-263
Hearmon RFS (1948) The elasticity of wood and plywood. Dept Sci Ind Res For Prod Res Spec
Report no 7. HMSO, London, 87 p
Hearmon RFS (1965) The assessment of wood properties by vibrations and high frequency
acoustic waves. 2nd Symp NDT, Washington State University, pp 49-66
Hearmon RFS, Burcham IN (1954) The dielectric properties of wood. For Prod Res Special Report
no 8, Dept of Sci and Industry, London, 25 p
Henneke EG, Jones TS (1979) Detection of damage in composite materials by vibrothermogra-
phy. ASTM Spec Tech Publ 696:83-95
Hennecke EG, Reifsnider KL, Stinchcomb WW (1979) Thermography - an NDE method for
damage detection. J Metals Sept:11-15
Herman GT (1980) Image reconstruction from projections. The fundamentals of computerized
tomography. Academic Press, London, 316 pp
Herman GT, Lent A, Rowland S (1973) Mathematics and applications: a report on the mathemat-
ical foundations and on applicability to real data of the algebraic reconstruction technique.
J Theor BioI 43:1-32
Hippel von A (1961) Dielectric materials and applications. Wiley, New York, 302 pp
Hoag M, McKimmy MD (1988) Direct scanning X-ray densitometry of thin wood sections. For
Prod J 38 1:23-26
Hodges DG, Anderson WC, McMillin CW (1990) The economic potential of CT scanners for hard-
wood sawmills. For Prod J 40 3:65-69
Horig H (1935) Theory of elasticity of anisotropic solids applied to wood (Anwendung der
Elastizitatstheorie anisotroper Korper auf Messungen an Holz). Ing Arch VI:8-14
Hounsfield GN (1972) A method and apparatus for examination of a body by radiation such as
X-ray or gamma radiation. Patent Specification 1283915, The Patent Office, London
Hoyle RJ (1961) A nondestructive test for stiffness of structural lumber. For Prod J 11 6:
251-254
308 References
Hsi E, Hossfeld R, Bryant RG (1977) NMR relaxation study of water absorbed on milled North
white cedar. J Colloid Interface Sci 62:389-395
Hutchins DA, Jansen DP, Edwards P (1993) Lamb wave tomography using noncontact transduc-
ers. Ultrasonics 31 :97 -103
Hutchins MT, Windsor CG (1987) Industrial applications. In: Sk6ld K, Price DL (eds) Neutron
scattering, vol 23, part C. Academic Press, Orlando, pp 405-407
Iancu L, Iancu G, Grimberg R, Lupu A (2000) Quantification of defects in wood by use of
ultrasonics in association with imagistic methods. 15th World Conf On NDT, Roma,
http://www.ndt.net/article/wcndtOO/paper
Ida N (1992) Microwave nondestructive testing. Kluwer, Amsterdam, 300 pp
Ida N, Wang JS (1996) Models for microwave nondestructive testing of materials. Mater Sci Forum
210-213:93-100
Jacquiot C, Trenard Y, Dirol D (1973) Atlas d'anatomie des bois des angiospermes. Tome 1. Centre
Technique du Bois, Paris, 150 pp
Jain VK, Sanyal SN (1996) A preliminary note on the attenuation of microwaves by few wood
species. Indian For 122,5:428-431
James WL (1959) A method for rapid measurement of the rate of decay of free vibrations. USDA,
FPL Bull no 2154, USDA, Madison, WI
James WL (1975) Dielectric properties of wood and hardboard: Variation with frequency, mois-
ture content and grain orientation USDA For Serv Res Pap FPL-245, Madison, WI
James WL (1989) Electrical properties. In: Schniewind AP (ed) Concise encyclopedia of wood and
wood-based materials. Pergamon Press, Oxford, pp 99-102
James WL, Hamill DW (1965) Dielectric properties of Douglas fir measured at microwave fre-
quencies. For Prod J 152:51-56
James WL, Yen YS, King, RJ (1985) A microwave method for measuring moisture content, density
and grain angle of wood. USDA For Serv Res Notes FPL-250, Madison, WI, 9 pp
Jartti P, Luukkala M (1977) Ultrasonic method for web speed measurements. Tappi J 60 11:167
Javadpour Z, Hughes D, Keating JG, Feeney FE, Evertsen JA (1996) Assessment of timber quality
using neural networks on CT images. Proc 10th Symp Nondestructive Testing of Wood Swiss
Federal Institute of Technology, Press Polytechniques et Universitaires Romandes, Lausanne,
pp 303-311
Jayne BA (1955) A nondestructive test of glue bond quality. For Prod J 5 5:294-301
Jayne BA (1959) Vibrational properties of wood as indices of quality. For Prod J 911:413-416
Jayne BA (1972) Theory and design of wood and fiber composite materials. Syracuse University
Press, Syracuse, NY, 418 pp
Johansson LG (1985) Experiences from using X-ray tomography, isotope-based nontomographic
measuring and NMR for testing logs before sawing 5th Symp Non destructive testing of Wood,
Washington State University, Pullman, pp 205-229
Kabir MF, Daud WM, Khalid K, Sidek HHA (1998) Dielectric and ultrasonic properties of rubber
wood. Effect of moisture content, grain direction and frequency. Holz Roh Werkst 56:223-227
Kabir MF, Schmoldt DL, Schafer ME (2000a) Ultrasonic detection of knots, cross grain and bark
pockets in wooden pallet parts. Proc World Conf Timber Eng 2000, Whistler, British Colum-
bia, Canada, pp 7521-7528
Kabir MF, Schmoldt DL, Schafer ME (2000b) Detection of defects in red oak deckboards by ultra-
sonic scanning. Proc 4th Int Conf Image Proccesing and Scanning of Wood, Mountain Lake
Resort, Virginia, pp 89-96
Kabir MF, Schmoldt DL, Schafer ME (2001) Roller transducer scanning of wooden pallet parts
for defect detection. In: Thompson DO, Chimenti DE (eds) Review of progress in quantitative
nondestructive evaluation, vol 20. Plenum Press, New York, pp 1218-1225
Kaestner AP, Baath L (2000) Microwave polarimetry based wood scanning. 12th Symposium on
Nondestructive Testing of Wood, Sopron, pp 349-356
Kiitzel R, Ridder HW, Habermehl A (1997) Investigations of ammonia stressed Scots pine trees
(Pinus sylvestris) with computed tomography in an immission area around a cattle farm.
Phyton 371:141-149
References 309
Kak AC (1979) Computerized tomography with X-ray emission and ultrasound sources. Proc
IEEE 67:1245-1272
Kak, AC Slaney M (1988) Principles of computerized tomographic imaging. IEEE Press, New York,
329 pp
Kanagawa Y, Hattori Y (1985) Nondestructive measurement of moisture distribution in wood with
a medical X-ray CT scanner. Part II. Changes in moisture distribution with drying. Mokuzai
Gakkaishi 31:983-989
Kanagawa Y, Furuyama Y, Hattori Y (1992) Nondestructive measurement of moisture diffusion
coefficient in wood drying. Drying Technoll0:1231-1248
Karsulovic JT, Leon LA, Dinator Mi (1999) The use of linear attenuation coefficients of gamma
radiation for detecting knots in Pinus radiata. For Prod J 49 2:73-76
Kasischke ES, Melack JM (1997) The use of imaging radars for ecological applications. Remote
Sensing Environ 59 2:141-156
Keller E, Thiercelin F (1975) The influence of big medulary rays on some properties of beech.
(Influence de gros rayons ligneux sur quelques proprieltels du bois de he3tre). Ann Sci For
322:113-129
Kessler R (1987) Applicability of imaging radar for classification of forest vegetation. Photo-
grammetria 414:221-232
Khan MA, Rizvi TZ, Naheed A, Zaheer MY, Iqbal MZ, Shoaib S (2000) Dielectric observed con-
sequences of microbial treatment. Holzforschung 54:335-339
Kharadly M (1985) Microwave diagnostic for stress rating of dimension lumber. Proc 5th Non-
destructive Testing of Wood, Washington State University, Pullman, pp 445-464
Kim YS, Singh AP (2000) Michromorphological characteristics of wood biodegradation in wet
environments: a review. IAWA J 21:135-155
King RJ (1978) Microwave electromagnetic nondestructive testing of wood. Proc 4th Nonde-
structive Testing of Wood, Washington State University, Pullman, pp 121-134
King RJ (2000) On-line industrial applications of microwave moisture sensor. In: Baltes H, Gopel
W, Hesse J (eds) Sensor update, vol 7. VCH Verlag, Weinheim, pp 109-170
King RJ, Basuel JC (1993) Measurement of basis weight and moisture content of composite boards
using microwaves. For Prod J 439:15-22
King RY, Yen YH (1981) Probing amplitude, phase and polarization of microwave field distribu-
tions in real time. IEEE Trans Microwave Theor Tech 29:1225-1231
Kinney JH, Johnson QC, Bonse U, Nichols MC, Saroyan RA, Nusshardt R, Pahl R, Brase JM (1988)
Three dimensional X-ray computed tomography in material science. MRS Bull 31 1:13-
17
Kittel C (1968) Introduction to solid state physics. Wiley, New York, 399 pp
Klein P, Vogel H (1993) Computertomography - a help for dendrochronology and wood identi-
fication? Proc 4th Int Conf on Nondestructive Testing of Works of Art, Berlin, pp 85-91
Kleinberg RL (1999) Nuclear magnetic resonance. In: Wang P-Z (ed) Methods in the physics of
porous media, vol 35. Academic Press, San Diego, pp 337-385
Kliger IR, Perstorper M, Johansson G, Pellicane PJ (1995) Quality of timber products from Norway
spruce. Part 3. Influence of spatial position and growth characteristics on bending stiffness
and strength. Wood Sci TechnoI29:397-410
Kline R, Sullivan C, Mignona RB, Delsanto PP (1996) A parallel processing algorithm for acoustic
tomography. In: Tortoli P, Masotti L (eds) Acoustical imaging. Plenum Press, New York, pp
561-566
Kline RA, Wang YQ (1992) A technique for ultrasonic tomography in anisotropic media. J Acoust
Soc Am 91:878-884
Kline RA, Wang YA, Mignogna RB, Delsanto PP (1994) A finite difference approach to acoustic
tomography in anisotropic media. J Nondestr Eval13:75-83
Koch M, Hunsche S, Schuacher P, Nuss MC, Feldmann J, Fromm J (1998) THz-imaging: a new
method for density mapping of wood. Wood Sci TechnoI32:421-427
Kollmann F (1951) Technology of wood and wood based composites (Technologie des Holzes
und der Holzwerkstoffe). Springer, Berlin Heidelberg New York, 1048 pp
310 References
Kollmann FFP, Cote WA (1968) Principles of wood science and technology. Springer, Berlin
Heidelberg New York, 592 pp
Kollmann FFP, Krech M (1960) Dynamic measurements of wood elastic properties and damping.
(Dynamische Messung der elastischen Holzeigenschaften und der Dampfung). Holz Roh
Werkst 18 2:41-45
Kostorz G, Lovesey SW (1979) Neutron scattering. General introduction. Treatise Mater Sci
TechnoI15:1-67
Kucera LJ (1990) Current use of the NMR on wood at the Swiss Federal Institute of Technology:
overview and outlook. 7th Int Symp on Nondestructive Testing of Wood, Washington State
University, Pullman, pp 71-72
Kuo PK, Feng Z1, Ahmed T, Favro LD, Thomas RL, Hartikainen J (1987) Parallel thermal wave
imaging using a vector lock-in video technique. In: Hess P, Pelzl J (Eds) Photo acoustic and
photothermal phenomena. Springer, Berlin Heidelberg New York, pp 415-418
Kuroda N, Suzuki Y (1996) Effects of morphological factors on the electrical behavior of wood
with the presence of free water. Current behavior from voltage reversals and tracheary ele-
ments. Mokuzai Gakkaishi 42:817-824
Lambert JB, Frye JS, Carriveau GW (1991) The structure of oriental lacquer by solid state NMR
spectroscopy. Archaeometry 33 87:93
Lambert JB, Shawl CE, Stearns JA (2000) Nuclear magnetic resonance in archeology. Chern Soc
Rev 29 3:175-182
Lanceleur P, de Belleval JM, Mercier N (1998) Synthetic tridimensional representation of
anisotropic materials. Acta Acust 84:1047-1054
Langwig JE, Meyer JA (1973) Ion migration in wood verified by neutron activation analysis. Wood
Sci 6 1:39-50
Lau SK, Almond DD, Milne JM (1991) A quantitative analysis of pulsed video thermography. NDT
Eval Int 24 4:195-202
Laufengerg TL (1986) Using gamma radiation to measure density gradient in reconstructed wood
products. For Prod J 36 (2):59-62
Laurence GR, Banner A (1980) The application of thermography for locating potential frost
pockets in forest cutovers. In: Proc of the 6th Canadian symposium on remote sensing,
Halifax, Nova Scotia. Canadian Aeronautics and Space Institute, Halifax, Canada, pp 369-374
Lebowitz C, Zoughi R, Spicer JWM, Osinder R (1996) Comparison of ultrasonic, microwave and
photothermal imaging of defective graphite-epoxy composite panels. Proc SPIE 2944:67-74
Lehmann E, Vontobel P, Niemz P (2001) Application of neutron radiography as method in the
analysis of wood. (Anwendung der Methods der Neutronenradiographie zur Analyse von
Holzeigenschaften). Holz Roh Werkst 59:463-471
Lehto A, Jaarinen J, Tiusanen M, Jokinen M, Luukkala M (1981) Amplitude and phase in thermal
wave imaging. Electr Lett 17:364-365
Lewa CJ (1991) Magnetic resonance imaging in the presence of mechanical waves. Spectrosc Lett
241:56-67
Lewa CJ (1992) NMR detection of elastic wave effects. In: Alippi A (ed) Acoustic sensing and
probing. World Science, Singapore, pp 435-438
Lewa CJ (1996) Elasto-magneitc resonance spectroscopy. Europhys Lett 35 1:73-76
Liezers M, Miller RM, Spillane DEM, Ryan TH (1985) Developments in thermal wave imaging.
Proc Int Conf Acoustic Emission and Photo acoustic spectroscopy and applications, London,
pp 65-69
Lin RT (1967) Review of the dielectric properties of wood and cellulose. For Prod J 177:61-66
Lin RT (1973) Wood as an orthotropic dielectric material. Wood Fiber 5 3:226-236
Lindberg H, Grahn 0, Lindgren A, Heligren C (1996) Watter diffusion through acrylate latex paint
films measured by computed tomography. Proc of 10th Nondestructive Testing of Wood Sym-
posium. Swiss Federal Institute of Technology, Press Poly techniques et Universitaires Roman-
des, Lausanne, 408 p
Lindegaard-Andersen A, Notea A, Bushlin Y, Rheinlander J (1990) Image-centred frame of refer-
ence in rotational film-based X-ray tomography. NDT Int 23 6:332-334
References 311
Lindgren LO (1985) On the relationship between density / moisture content in wood and X-ray
attenuation in computer tomography. Proc 5th Symp Nondestructive Testing of Wood;
Washington State University, Eng Publications, Pullman, pp193-203
Lindgren 0 (1991a) Medical CAT-scanning: X-ray absorption coefficients, CT-numbers and their
relation to wood density. Wood Sci TechnoI25:341-349
Lindgren 0 (1991b) The accuracy of medical CT-scan images for nondestructive density mea-
surements in small volume elements within solid wood. Wood Sci Technol 25:425-432
Lindgren 0 (1994) NMR for nondestructive wood moisture content measurement. First European
Symp on NDE of Wood, University of Sopron, Sopron, pp 124-129
Lindgren 0, Davis J, Wellq P, Shadbolt P (1992) Nondestructive wood density distribution mea-
surements using computed tomography. Holz Roh Werkst 50:295-299
Liu CJ, Olson JR, Tian Y, Shen Q (1988) Theoretical wood densitometry: I. Mass attenuation equa-
tions and wood density models. Wood Fiber Sci 20:22-33
Loos WE (1965) Determining moisture content and density of wood by nuclear radiation tech-
niques. For Prod J 153:102-106
Loos WE (1965) Neutron activation of Engelmann spruce. For Prod J 154:178
Luong PM (1995) Infrared scanning of failure process in wood. SPIE 2473:298-309
Luong PM (1996) Infrared thermography of damage in wood. 10th Symp NDT of Wood, Press
Univ Romandes, Lausanne, pp 175-185
Luong PM (1998) Fatigue limit evaluation of metals using an infrared thermographic technique.
Mech Mater 28:155-163
Luukkala M, Heikilla P, Surakka Y (1971) Plate-wave resonance. A contactless test method. Ultra-
sonics 9:201-208
Lyamshev ML, Stanullo 1, Busse G (1995) Thermo acoustic vibrometry. Remote in situ monitor-
ing of ceramic sintering. Materialpriifung 37 112:22-24
Lynnworth LC (1989) Ultrasonic measurements for process control. Theory, techniques, applica-
tions. Academic Press, Boston, 694 pp
Lynnworth LC, Magory V (1999) Industrial process control sensors and systems. In: Papadakis
EP (ed) Ultrasonic instruments and devices. Academic Press, San Diego, pp 276-470
Lynnworth LC, Nguyen TH, Smart CS, Khrakovsky OK (1997) Acoustically isolated paired air
transducers for 50, 100,200 or 500kHz applications. IEEE Trans UFFC 44 5:1087-1100
Lytle RJ, Dines K (1980) Iterative ray tracing between boreholes for underground image recon-
struction IEEE Trans Geosci Remote Sens GE 18:234-240
MacFali JS, Johnson GA, Kremer PJ (1990) Observation of water-depletion region surround-
ing loblolly pine roots by magnetic resonance imaging. Proc Natl Acad Sci USA 87:1203-
1207
MacGregor RP, Peemoeller H, Schneider MH, Sharp AR (1983) Anisotropic diffusion of water in
wood. J Appl Polymer Sci Polymer Symp 37:901-909
Madsen B (1994) Radiological density scanning - a portable gamma camera based on back-
scatter tomography. Proc 9th Symp Nondestructive Testing of Wood, Washington State
University, Madison, pp 131-137
Makoviny I (1988) The anisotropy of wood dielectric constants as a function of anatomical
structure. (Zur Anisotropie der Dielektrizitatskonstanten des Holzes in grundlegenden
anatomischen Richtungen). Holztechnologie 29 4:210-213
Malan FS, Marais PG (1992) Some notes on the direct gamma ray densitometry of wood. Holz-
forschung 46:91-97
Maloney TM (1992) Composites for the future. Invited paper. Proc IUFRO Conference, Division
5, Nancy, 605p p
Maloney TM (1996) The family of wood composite materials. For Prod J 46 2:19-26
Mandelis A (2000) Diffusion waves and their uses. Physics Today 53 No 8 Part 1:29-34
Mansfield P, Morris PG (1982) NMR imaging in bio-medicine. In: Waugh J (ed) Advanced mag-
netic. Resonance Suppl no 2. Academic Press, New York
Marcinko JJ, Devathala S, Rinaldi PL, Bao S (1998) Investigating the molecular and bulk dynam-
ics of PMDIIwood and UF/wood composites. For Prod J 48 6:81-84
312 References
Marcinko JJ, Rinaldi PL, Bao S (1999) Exploring the physicochemical nature of PMDl/wood struc-
tural composite adhesion. For Prod J 49 5:75-78
Marok M, Kundela J, Cunderlik I (1996) Identification of reaction beech wood by X-ray computed
tomography. Holz Roh Werkst 54:97-98
Martin P, Collet R, Barthelemy P, Roussy G (1987) Evaluation of wood characteristics: internal
scanning of the material by microwave. Wood Sci TechnoI21:361-371
Martinis R (2002) Nondestructive techniques for decay diagnosis on standing trees. (Analisi di
tecniche non destruttive per la diagnosi di carie su alberi in piedi). Thesis University of
Florence, Florence, Italy, 300 pp
Masuch G, Franz JT, Marsmann H, Gross D, Kettrup A (1991) 'H NMR micro-imaging and
correlated SEM studies of spruce needles from healthy and declined forest sites. Int J Environ
Anal Chern 45:179-191
Masuda M, Takahashi S (1998) Change of thermal images of lumber including knots, finger joints
and metal connectors. Bull Kyoto Univ For (Kyoto Daigaku Nogaku Bu Enshurin Hokoku)
69:114-128
Masuda M, Takahashi S (1999) Thermographical detection of knots and other defects. Proc
INCEUPT '99 For Prod Assoc ROC Bull 16:181-187
Masuda M, Takahashi S (2000) Thermographical detection of defects of lumber and glued
joints. 12th Symp NDT of Wood, University of Western Hungary, Sopron, pp 421-
430
Masuda M, Fujimoto K, Takino S, Sadoh T (1995) Change of thermal images of timber including
knots under repeated bending. Bull Kyoto Univ For 67:167-173
McDonald KA (1978) Lumber defect detection by ultrasonics. Forest Prod Lab Res Paper FPL 311,
USDA, Madison, WI, 25 pp
McGaughey WJ, Young RP (1990) Comparation of ART, SIRT, Least-Squares, and SVD two-
dimensional tomographic inversions of field data. 60th Annual Meeting of Soc Exploration
Geophysicists SEG, pp 74-77
McKeon JCP, Hinders MK (1999) Parallel projection and crosshole Lamb contact scanning tomog-
raphy. JASA 106 5:2568-2577
McLauchlan TA, Norton JA, Kusec DJ (1973) Slope of grain indicator. For Prod J 23 5:50-55
McMillin CW (1982) Application of automatic image analysis to wood science. Wood Sci 14:
97-105
Meiboom S, Gill D (1958) Modified spin-echo method for measuring nuclear relaxation times.
Rev Sci Instrum 29:688-691
Menon RS, MacKay AL, Flibotte S, Hailey JRT, Blook M, Burgess AC, Swanson JS (1987) A NMR
determination of the physiological water distribution in wood during drying. J Appl Polymer
Sci 33:1146-1155
Menon RS, MacKay AL, Flibotte S, Hailey JRT (1989) Quantitative separation of NMR images of
water in wood on the basis of T(2). J Magn Reson 82 205-210
Millard P, Chudek JA (1993) Imaging the vascular continuity of Prunus avium petioles during leaf
senescence using NMR spectroscopy. J Exp Bot 44:599-603
Miller WH (1988) Design and implementation of a portable computerized axial tomography
system for field. Nucl Instrum Meth Phys Res 270:590-597
Moksin MM (1993) Improvement of an opto-thermal technique for wood. Wood Sci Technol28:
53-58
Moschler WW Jr, Dougal EF (1988) Calibration procedure for a direct scanning densitometer
using gamma radiation. Wood Fiber Sci 20:297-303
Moschler WW Jr, Winistorfer PM (1990) Direct scanning densitometry: an effect of sampling het-
erogeneity and aperture area. Wood Fiber Sci 22:31-38
Mothe F, Duchanois G, Zannier B, Leban JM (1998) Analyse microdensitometrique appliquee au
bois. Ann Sci For 55:301-313
Murata K, Sadoh T (1994) Heat absorption and transfer in softwoods and their knot surfaces.
Mokuzai Gakkaishi 40:1180-1184
Musial MW (1988) A nondestructive method for determining the degree of flake orientation in
OSB. Wood Sci TechnoI22:371-378
References 313
Naito S, Sawada Y, Fujii Y, Okumura S (1998) Thermography of wood specimens in static bending
test. Kyoto Daigaku Nogaku Bu Enshurin Hokoku 69:104-113
Naito S, Fujii Y, Sawada Y, Okumura S (2000) Thermographic measurement of slope of grain using
the thermal anisotropy of wood. Mokuzai Gakkaishi 46:320-325
Nakanishi TM (1994) The water image of plants using neutron radiography. Chern Regul Plants
292:182-190
Nakanishi TM, Matsubayashi M (1997) Nondestructive water imaging by neutron beam analysis
in living plants. J Plant PhysioI151:442-445
Nakanishi TM, Watanabe S (1995) Water distribution inside the woods by neutron radiography.
Bull Tokyo Univ For 93:13-19
Nakanishi TM, Karakama I, Sakura T, Matsubayashi M (1998a) Moisture imaging of a camphor
tree by neutron beam. Radioisotopes 47:387-391
Nakanishi TM, Okano T, Karakama I, Ishihara T, Matsubayashi M (1998b) Three dimensional
imaging of moisture in wood disk by neutron beam during drying process. Holzforschung
52:673-676
Nanassy AJ (1970) Overlapping of dielectric relaxation spectra in oven -dry birch at temperatures
from 20 to 100 degrees C. Wood Sci TechnoI4:104-121
Nanassy AJ (1972) Dielectric measurements of moist wood in a sealed system. Wood Sci Technol
6:67-77
Nanassy AJ (1973) Use of wide line NMR for measurement of moisture content in wood. Wood
Sci 5:187-193
Nanassy AJ (1974) Water sorption in green and remoistured wood studied by the broad-line com-
ponent of the wide line NMR spectra. Wood Sci 7:61-68
Natterer J, Sandoz JL, Rey M (2000) Bois matelriau et struture bois. Press Polytheqnique et
Universitaire Romande, Lausanne, 488 pp
Neuenschwander J, Niemz P, Kucera LJ (1997) Studies for visualizing wood defects using ultra-
sonic in reflection and transmission mode. Holz Roh Werkst 55:339-340
Newmann RH (1992) NMR study of spatial relationship between components in wood cell walls.
Holzforschung 46:207-210
Nicklow RM (1979) Phonons and defects. Treatise Mater Sci TechnoI15:191-225
Nicolotti G, Miglietta P (1998) Using high-technology instruments to assess defects in trees.
J Arbor 24 6:297-302
Niemienen AOK, Koenig JL (1988) NMR imaging - a promising new adhesive evaluation
technique. J Adhesion Sci TechnoI2:407-414
Niemz P (1989) Use of online moisture-measuring instruments in the woodworking industry: lit-
erature review (Einsatz von on-line-Feuchtemessgeraten in der holzverarbeitenden Industrie
- eine Literaturubersicht). Holztechnologie 30 1:9-13
Niemz P, Sander D (1989) Mesuring techniques in wood industrie. Prozessmesstechnick in der
Holzindustrie. VEB Fachbuchverlag, Leipzig, 288 pp
Niemz P, Kucera LJ, Flisch A, Blaser E (1997) Anwendung der Computertomographie an Holz.
(Computerized tomography of wood). Holz Roh Werkst 55:279-280
Niemz P, Bodmer HC, Kucera LJ, Ridder HW, Habermehl A, Wyss P, Zurcher E, Holdenrieder 0
(1998a) Eignung verschiedener Diagnosemethoden zur Erkennung von Stammfliulen bei
Fichte (Suitability of different methods for decay detection in Norway spruce). Schweiz.Z
Forstwes 149:615-630
Niemz P, Kucera LJ, Ridder HW, Habermehl A, Flisch A (1998b) Durchblick mit Computertomo-
graphie (Transversal view with computer tomography). Wald Holz 9:7-10
Niemz P, Kucera LJ, Schob M, Scheffer M (1999) Experiments for defect detection in wood with
ultrasound (Experimentelle Untersuchungen zur Erkennung in Holz mit Ultraschal). Holz
Roh Werkst 57:96-102
Niemz P, Kucera LJ, Lehmann E, Vontobel P, Hanschke S (2000a) Investigation of wooden corner
connection by neutron radiography methods in order to evaluate the ingress behavior of
water. 12th Symp Nondestructive Testing of Wood, Sopron, p 463
Niemz P, Lehmann E, Vontobe1 P (2000b) Application of neutron radiography for investigation of
wood. 12 th Symp Nondestructive Testing of Wood, Sopron, p 462
314 References
Niemz P, Lehmann E, Vontobel P, Haller P, Hanschke S (2000c) Investigation using neutron radi-
ography for evaluations of moisture ingress into corner connections of wood. 12th Symp on
Wood, University of Western Hungary, Sopron, pp 489-497
Nishino Y, Norimoto M (1990) Structure and anisotropy of dielectric constant in hardwood. Wood
Resh Tech Notes Kyoto Univ 26:78-90
Norimoto M (1976) Dielectric properties of wood. Wood Res Kyoto 59/60:106-152
Norimoto M, Yamada (1972) The dielectric properties of wood. VI. On the dielectric properties
of chemical constituents of wood and the dielectric anisotropy of wood. Wood Res Bull Wood
Res Inst Kyoto Univ:31-43
Norimoto M, Zhao G (1993) Dielectric relaxation of water adsorbed on wood II. Mokuzai
Gakkaishi 39:249-257
Norimoto M, Hayashi S, Yamada T (1978) Anisotropy of dielectric constant in coniferous wood.
Holzforschung 32:167-172
Norton SJ, Linzer M (1980) Ultrasonic reflectivity imaging in three dimensions: exact inverse
scattering solutions for plane, cylindrical and circular apertures. IEEE Trans Biomed Eng
28:202-220
Nystrom J, Earl Kline D (2000) Automatic classification of compression wood in green Southern
yellow pine. Wood Fiber Sci 32:301-310
Obataya E, Yokoyama M, Norimoto M (1996) Mechanical and dielectric relaxations of wood in a
low temperature range. I. Relaxations due to methylol groups and adsorged water. Mokuzai
Gakkaishi 42:243-249
Oja J, Temnerud E (1999) The appearance of resin pockets in CT-images of Norway spruce (Picea
Abies (L) Karst). Holz Roh Werkst 57:400-406
Oja J, Grundberg S, Gronlund A (2000) Predicting the strength of sawn products by X-ray scan-
ning of logs: a preliminary study. Wood Fiber Sci 32:203-208
Okumura S, Suzuki R, Fujii Y (1996) Preliminary study on thermography of wood under com-
pression load. Kyoto Daigaku Nogaku Bu Enshurin Hokoku 68:161-169
Olek W, Guzenda R, Baranowska HM, Olszewski J (1994) On possible application of nuclear mag-
netic relaxation measurements to investigation of water penetration into wood. First Euro-
pean Symp on NDE of Wood, University of Sopron, Sopron, pp 130-137
Olek W, Guzenda R, Weres J (2000) Computed-aided estimation of thermal conductivity of Scots
pine wood. 12th Symp NDT of Wood, University of Western Hungary, Sopron, pp 431-438
Olson JR,Arganbright DG (1981) Prediction of mass attenuation coefficients of wood. Wood Sci
14:86-90
Olson JR, Liu CJ, Tian Y, Shen Q (1988) Theoretical wood densitometry: II. optimal X-ray energy
for wood density measurement. Wood Fiber Sci 20:187-196
Olson JR, Chang SJ, Wang PC (1990) Nuclear magnetic resonance imaging: a noninvasive analy-
sis of moisture distributions in white oak lumber. Can J For Res 20:586-591
Onoe M, Tsao JW, Tamada H, Nakamura H, Kogure J, Kawamura H, Yoshimatsu M (1984) Com-
puted tomography for measuring the annual rings on a live tree. Nuclear Instruments and
Methods in Physics Research 221:213-220
Pang S (1996) Moisture content gradient in a softwood board during drying: simulation from
2-D model and measurement. Wood Sci TechnoI30:165-178
Pang S, Wiberg P (1998) Model predicted and CT scanned moisture distributed in a Pinus radiata
board during drying. Holz Roh Werkst 56:9-14
Pang S, Keey RB, Laugrish TAG, Walker JCF (1994) Air flow reversals in high temperature kiln
drying of Pinus radiata boards:l: drying of a single board. N Z J For Sci 24:83-103
Papadakis EP (1999) Ultrasonic instruments and devices. Reference for modern instrumentation,
techniques and technology. Academic Press, San Diego, 809 pp
Passialis C (1997) Physico-chemical characteristics of waterlogged archeological wood. Holz-
forschung 51:11-113
Pearce RB, Sumer S, Doran SJ, Carpenter TA, Hall LD (1994) Noninvasive imaging of fungal col-
onization and host response in the living sapwood of sycamore (Acer pseudoplaranus) using
nuclear magnetic resonance. Physiol Plant Pathol 45:359-384
References 315
Pearce RB, Fisher BJ, Carpanter TA, Hall LD (1997) Water distribution in fungal lesions in the
wood of sycamore (Acer pseudoplatanus), determined gravimetrically and using nuclear mag-
netic resonance imaging. New Phytol 135:675-688
Pei J, Yousuf MI, Degertrkin FL, Honein BV, Kuri-Yakub BT (1995) Lamb wave tomography and
its application in pipe erosion/corrosion monitoring. Proc IEEE Ultrasonics Symp 1995,
Seattle, Washington, pp 795-798
Pellerin RF (1965) A vibrational approach to nondestructive testing of structurallumber. For Prod
J 153:93-101
Peters RA (1995) A new algorithm for image noise reduction using mathematical morphology.
IEEE Trans Med Imag 4:554-568
Peterson KR (1993) The role of nondestructive evaluation in assuring the wise use of our
timber resource. 9th Symp NDT of Wood, Washington State University Press, Madison,
pp 7-9
Peyskens E, Pourq M de, Stevens M, Schalck J (1984) Dielectric properties of softwood species at
microwave frequencies. Wood Sci TechnoI18:267-280
Pfeffer PE, Gerasimowicz WV (1989) Nuclear magnetic resonance in agriculture. CRC Press, Boca
Raton, 441 pp
Polge H (1966) L'analyse densitomeltrique de cliches radiographiques: Une nouvelle methods de
determination de la structure du bois. Ann Sci For 20 4:530-581
Polge H (1978) Fifteen years of wood radiation densitometry. Wood Sci TechnoI12:187-196
Poliszko S, Hoffmann G (1985) Dielectric behavior of wood-polystyrene composite. J Appl
Polymer Sci 30:799-804
Price DL, Sk6ld K (1986) Introduction to neutron scattering. In: Sk6ld K, Price DL (eds) Neutron
scattering, vol 23, part A. Academic Press, Orlando, pp 1-98
Purslow DF (1971) The use of microwave moisture meter for studying moisture changes in
timber. J Inst Wood Sci 5 4:40-46
Puttick KE (1987) Thermal NDT methods. In: Summers cales J (ed) Nondestructive testing of
fibre-reinforced plastics composites. Elsevier, London, pp 65-103
Quick J], Hailey JRT, Mac Kay AL (1990) Radial moisture profiles of cedar sapwood during drying:
a proton magnetic resonance study. Wood Fiber Sci. 22:404-412
Quin F Jr, Steele PH, Shmulsky R (1998) Locating knots in wood with an infrared detector system.
For Prod J 48 10:80-84
Radon J (1917) Uber die Bestimmung von Funktionen durch ihre Integralwerte langs gewisser
Manningfaltigkeiten (On the determination of functions from their integrals along certain
mainfolds). Ber Sach Akad Wiss 29:262-277
Rafalski J (1966) On the dielectric properties of compress red beech (Uber die dielekkyrischen
Eigenschaften unterschiedich verdiichteten Rotbuchen vollholzes). Holztechnologie 7 2:118-
122
Ranta L, May HA (1978) Measuring the density profile of particleboard by gamma radiation. Holz
Roh Werkst 36:467-474
Rantala J, Wu D, Busse G (1996) Vibrothermography applied for nondestructive evaluation of
polymer materials. Mater Sci Forum 210-213:433-438
Raschi A, Tognetti R, Ridder HW, Beres C (1995) Water in the stems of sesilis oak (Quercus
petraea) by computer tomography with concurrent measurements of sap velocity and ultra-
sound emission. Plant Cell Environ 18:545-554
Reimers P, GilboyWB, Goebbels J (1984) Recent development in the industrial application of com-
puterized tomography with ionizing radiation. NDT Int 174:197-207
Reynolds WN, Wells GM (1984) Video compatible thermography. Br J NDT 26 1:40-44
Rice RW, Steele PH, Kumar L (1992): Detecting knots and voids in lumber with dielectric sensors.
Ind Metrol2 3/4:309-315
Richards JA (1990) Radar backscatter modeling of forests: a review of current trends. Int J Remote
Sens 11:1299-1312
Rinn F (1991) Three dimensional impulse tomograph for examination of trees and timber. Engi-
neering and Distribution, Heidelberg, 4 pp
316 References
Robertson MB, Packer KJ (1999) Diffusion of D20 in archaeological wood measured by I-D NMR
profiles. Appl Magn Reson 17:49-64
Rosencwaig A, Busse G (1980) High resolution photoacoustic thermal wave microscopy
Appl.Phys.Lett 36 : 725-727
Rosencwaig A (1983) Photo acoustics in material science. In: Herman H (ed) Treatise on
material science and technology, vol 198. Academic Press, NewYork, pp 67-117
Ross RJ, Pellerin RF (1991) Nondestructive evaluation of wood past, present and future. In:
Ruud CO, Green RE (eds) Nondestructive characterization of material, vol IV, Plenum Press,
New York, pp 59-64
Ross RJ, Pellerin RF (1994) Nondestructive testing for assessing wood members in structures. A
review. Gen Tech Rep FPL-GTR-70, USDA, Forest Product Laboratory, Madison, WI, 40 pp
Rothwell WP, Holecek DR (1984) NMR imaging: study of fluid absorption by polymer compos-
ites. J Polymer Sci Polymer Lett Ed 22:241-247
Roux S (1998) Scales. In: Frantziskonis GN (ed) Proc PROBAMAT - 21st Century: probability and
materials. Test, models and applications for the 21th century. Kluwer, Dordrecht, pp 573-
577
Rust S (2000) A new tomographic device for the nondestructive testing of trees. 12th Symp on
Nondestructive Testing of Wood, University of Western Hungary, Sopron, pp 233-237
Sadoh T, Murata K (1993) Detection of knots in hinoki and karamatsu lumber by thermography.
Mokuzai Gakkaishi 39:13-18
Sahimi M, Arbabi S (1993) Mechanics of disordered solids II. Phys Rev B 47:703-712
Salerno A, Wu D, Busse G, Rantala J (1997) Thermographic inspection with ultrasonic excitation
In: Thompson DO, Chimenti DE (eds) Review of progress in quantitative nondestructive eval-
uation, vol 16. Plenum Press, New York, pp 345-352
Salerno A, Dillenz A, Wu D, Rantala J, Busse G (1998) Progress in ultrasonic lockin thermogra-
phy. In: Balageas D, Busse G, Carlomagno GM (eds) Quantitative infrared thermography.
QIRT 98, Lodz, pp 164-160
Samson M (1988) Traverse scanning for automatic detection of general slope of grain in lumber.
For Prod J 38 7/8:33-38
Sasov A, van Dyck D (1998) Desktop X-ray microscopy and microtomography. J Microsc 191 pt
2:151-158
Schau D (1984) Fourier transform NMR. Elsevier, Amsterdam, 344 pp
Schechter RS, Chaskelis HH, Mignogna RB, Delsanto PP (1994) Real time parallel computation
and visualization of ultrasonic pulses in solids. Science 265:1188-1196
Schechter RS, Mignona RB, Delsanto PP (1996) Ultrasonic tomography using curved ray paths
obtained by wave propagation simulations on a massively parallel computer. J Acoust Soc Am
104:2103-2111
Schienwind AP (1990) Physical and mechanical properties of archeological wood. In: Rowell RM,
Barbour RJ (eds) Archeological wood properties, chemistry and preservation. American
Chemical Soc, Washington, DC, pp 87-109
Schirone A, Lo Monaco A (1988) L'uso dei raggi X in xililogia con particolare riferimento alle
indagini dendroclimatologiche (The utilization of X-ray for dendroclimatology). Economia
Montana (Linea Ecologica) 5:47-54
Schmidt RG, Frazier CE (2000) Solid state NMR analysis of adhesive bondlines in pilot scale flake-
boards. Wood Fiber Sci 32:419-425
Schmidt T (1978) Scanning computing methods for measuring knots and other defects in lumber
and veneer. Proc 4th Nondestructive Testing of Wood, Washington State University, Pullman,
pp 23-25
Schmoldt DL (1996) CT imaging, data reduction, and visualization of hardwood logs. In: Meyer
DA (ed) Proc 24th Annual Hardwood Symposium, National Lumber Assoc Cashiers, National
Hardwood Lumber Assoc, Cashiers, North Carolina, pp 69-80
Schmoldt DL, Zhu D, Conners RW (1993) Nondestructive evaluation of hardwood logs using auto-
matic interpretation of CT images. In: Thompson DO, Chimenti DE (eds) Review of progress
in quantitative nondestructive evaluation, vol 12. Plenum Press, New York, pp 2257-2264
References 317
Schmoldt DL, Li P, Araman PA (1996a) Iterative simulation of hardwood log veneer slicing using
CT images. For Prod J 46 4:41-47
Schmoldt DL, Nelson RM, Ross RJ (1996b) Ultrasonic defect detection in wooden pallet parts for
quality sorting. Proc SPIE 2944:285-295
Schmoldt DL, Pei Li A, Abbot Lyn A (1997) Machine vision using artificial neural networks with
local 3-D neighborhoods. Comput Electr Agric 16:255-27l
Schmoldt DL, Occena LG, Lynn Abbott A, Gupta NK (1999) Nondestructive evaluation of hard-
wood logs: CT scanning, machine vision and data utilization. Nondestr Test Eval 15:279-309
Schmoldt DL, He J, Lynn Abbott A (2000) Automated labeling of log features in CT imagery of
multiple hardwood species. Wood Fiber Sci 32:287-300
Schneider R, Netzelmann U, Kroning M (1996) Characterization of green ceramics by microwaves
and ultrasound. Mater Sci Forum 210-213:77-84
Schniewind AP (1981) Concise encyclopedia of wood and wood-based materials. Pergamon Press,
Oxford, 354 pp
Schomberg H (1982) Nonlinear image reconstruction from ultrasonic time to flight projections.
Acoust Imaging 10:381-396
Schulte M, FrUhwald A, Welling H, Kruse K, Broker FW (1995). Nondestructive in-line quality
control in board production. Tappi Proc of 1995 European plastic laminates forum "The
leading edge", Heidelberg, pp 7l-79
Schultz V, Ward Whicker F (1982) Radioecological techniques. Plenum Press, New York, 298 pp
Senft JF, Suddarth SK, Angleton RD (1962) A new approach to stress grading of lumber. For Prod
J 124:183-186
Sepulveda P, Gonlund A (2000) Measurement of spiral grain with computed tomography com-
pared with pattern on boards. Proc 12th Symp Nondestructive Testing of Wood, Sopron, pp
239-244
Sharp AR, Riggin MT, Kaiser R (1978) Determination of moisture content of wood by pulsed
nuclear magnetic resonance. Wood Fiber 10:74-81
Siau JF (1984) Transport process in wood. Springer, Berlin Heidelberg New York, 245 pp
Siau JF (1995) Wood: influence of moisture on physical properties. Department of Wood Science
and Forest Products, Virginia Polytechnic Institute and State University, Blacksburg, Virginia,
227 pp
Siau JF, Meyer JA (1973) Neutron activation analysis of pentachlorophenol in treated wood. Wood
Sci 6 1:19-21
Sieber AJ (1985) Forest signatures in imaging and nonimaging microwave scatterometer data.
ESA J 9 4:431-448
Skaar C (1948) The dielectric properties of wood at several radio frequencies. Tech Pub no 69.
NY State College of Forestry, Syracuse
Skaar C (1988) Wood in water relations. Springer, Berlin Heidelberg New York, 283 pp
Skatter S (1998) Determination of cross-sectional shape of softwood logs from three X-ray pro-
jections using an elliptical model. Holz Roh Werkst 56:179-186
Slichter CP (1978) Principles of magnetic resonance. Springer, Berlin Heidelberg New York, 397 pp
Schniewind AP (1990) Physical and mechanical properties of archaeological wood. In: Rowell RM,
Barbour RJ (eds) Archaeological wood: properties, chemistry and preservation. Advances in
Chemistry Series no 225, pp 87-109
Sobue N (1993) Nondestructive characterization of wood. Mokuzai Gakkaishi 39:973-979
Socco V, Martinis R, Sambuelli L, Comino E, Nicolotti G (2000) Open problems concerning ultra-
sonic tomography for wood decays diagnosis. 12th Symp NDT of Wood, University of Western
Hungary, Sopron, p 468
Socco V, Sambuelli L, Martinis R, Nicolotti G, Comino E (2002) Feasability of ultrasonic tomog-
raphy for NDT of decay on living trees. Res Nondestr Eval (in press)
Southon TE, Mattsson A, Jones RA (1992) NMR imaging of roots: effects after root freezing of
containerized conifer seedlings. Physiol Plant 86:329-334
Steele PH, Cooper JE (2000) Estimating lumber strength with radio frequency scanning. 12th
Symp Nondestructive Testing of Wood, Sopron, pp 343-348
318 References
Steele PH, Harless TEG, Wagner FG, Taylor FW (1989) Potential dollar increase from internal log
information. Proc 7th Symp on Nondestructive Testing of Wood, Washington State University,
Pulman, pp 241-250
Steele PH, Neal SC, McDonald KA, Cramer SM (1992) The slope of grain indicator for defect detec-
tion in unplaned hardwood lumber. For Prod J 41 1:15-20
Steele PH, Quin F Jr, Cooper E (1998) Infrared detection of knots in southern yellow pine lumber
heated with radiant heat. 11 th Symp NDT of Wood, Washington State University, Pullman, pp
57-62
Steele PH, Patton MD, Cooper E (2000a) Factor influencing differential thermal response of knots
and clear wood. 12th Symp NDT of Wood, University of Western Hungary, Sopron, pp 403-412
Steele PH, Patton MD, Cooper E (2000b) A thermal recognition system to identify knots in wood.
Proc wood technology. Clinic and Shows, Portland, Oregon, pp 255-266
Steele PH, Lalit Kumar, Shmulsky R (2000c) Differentiation of knots, distorted grain and clear
wood by radio frequency scanning. For Prod J 50 3:59-62
Sugimori M, Lam F (1999) Macro-void distribution analysis in strand-based wood composites
using an X-ray computer tomography technique. J Wood Sci Jpn Wood Res Soc 45:245-257
Sulivan C, Kline R, Mignogna RB, Delsanto PP (1996) A parallel processing approach to acoustic
tomograph. J Acoust Soc Am 99:2142-2147
Suryoatmono B, Cramer S, Shi Y, McDonald KA (1994) Within-board lumber density variations
from digital X-ray images. 9th Symp Nondestructive Testing of Wood, Washington State
University. Eng Publications, Madison, pp 168-175
Swanson JS, Hailey JRT (1987) Scanning and imaging techniques for assessing decay and wood
quality in logs and standing trees. Proc 6th Symp Nondestructive Testing of Wood, Washing-
ton State University. Eng Publications, Pullman, pp 83-94
Szendrodi L, Habermehl A, Ridder HW (1994) Computer tomographic investigation of stand-
ing trees in Hungary. 1st European Symp on Nondestructive Evaluation of Wood, Sopron, pp
503-511
Szymani R, McDonald KA (1981) Defect detection in lumber: state of the art. For Prod J 31
11:34-44
Talpe J (1971) Theory of experiments in paramagnetic resonance. Pergamon Press, Oxford
Tan HS (1981) Microwave measurements and modeling of the permittivity of tropical vegetation
samples. Appl Phys 25:351-355
Tanaka T (1994) Thermographic detection of deteriorated locations and nondestructive evalua-
tion of strength of biodeteriorated wood. 9th Symp NDT of Wood, Washington State Univer-
sity, Pullman, pp 78-83
Tanaka T, Divos F (2000) Thermographic inspection of wood. 12th Symp NDT of Wood, Univer-
sity of Western Hungary, Sopron, pp 439-447
Tanaka T, Norimoto M, Yamada T (1975a) Anisotropy of dielectric constant of wood. J Soc Mater
Sci 24 264:867-872
Tanaka T, Norimoto M, Yamada T (1975b) Dielectric properties and structure of wood I. Mokuzai
Gakkaishi 21:129-134
Taylor FW, Wagner FG Jr, McMillin CW, Lon Morgan I, Hopkins FF (1984) Locating knots by
industrial tomography - a feasibility study. For Prod J 34 5:42-46
Thomas RL, Favro LD, Kuo PK,Ahmed T, Han X, Wang L, Wang X, Shepard SM (1995) Pulse echo
thermal wave imaging for nondestructive evaluation. Proc 15 th Int Confrerence Acoustics,
Trondheim, pp 433-436
Thompson F (1989) Moisture measurements using microwaves. Meas Control 22:210-215
Thompson F (1996) Non destructive moisture measurement using microwaves. Mater Sci Forum
210-213:85-92
Tice D, Euskirchen J (1976) Faster and more effective forest fire surveillance with live IR imaging
superimposed on visible terrain view. Proc XVlth IUFRO World Congress, Oslo, pp 53-59
Tiitta M, Olkkonen H, Lappalainen T, Kanko T (1996) Density measurements of particleboard
veneer and wood specimens by narrow beam gamma absorption technique. 10th Symp Non-
References 319
destructive Testing of Wood Swiss Federal Institute of Technology, Press Univ Romande,
Lausanne, pp 187-200
Tinelli A, Catena G (1991) Injuries on ornamental trees in Castelporziano Indagine sulle piante
monumentali della tenuta di Castelporziano. Monti Boschi 5:9-13
Tiuri M, Jokela K, Heikkela S (1980) Microwave instrument for accurate moisture and density
measurement of timber. J Microwave Power 15:251-254
Tognetti R, Raschi A, Beres C, Fenyvesi A, Ridder HW (1996) Comparison of sap flow, cavitation
and water status of Quercus petraea and Quercus cerris trees with special reference to com-
puter tomography. Plant Cell Environ 19:928-938
Tomikawa Y, Iwase Y, Arita K, Yamada H (1985) Nondestructive inspection of rotted or termite
damaged wooden poles by ultrasound. Proc 5th Symposium on Ultrasonic Electronics, Tokyo
1984. Jpn J Appl Phys 24 1985 Suppl24 1:187-189
Tomikawa Y, Iwase Y, Arita K, Yamada H (1986) Nondestructive inspection of wooden pole using
ultrasonic computed tomography. IEEE Trans UFFC 33 4:354-358
Torgovnikov GI (1990) Dielektrische Eigenschaften von absolut trockenem Holz und Holzstoff
(Dielectric properties of oven dried wood). Holztechnologie 30 1:9-11
Torgovnikov GI (1993) Dielectric properties of wood and wood based materials. Springer, Berlin
Heidelberg New York, 196 pp
Tremblay C (1995) Longitudinal and radial variation of slope of grain in black spruce lumber.
For Prod J 45 1:79-83
Troughton GE, Clarke MR (1987) Development of a new method to measure moisture content in
unseasoned veneer and lumber. For Prod J 37 1:13-19
Tsai WH (1985) Moment-preserving thresholding: a new approach. Comput Vision Graph 29:
377-393
Tsuchikawa S, Takahashi T, Tsutsumi S (2000) Nondestructive measurement of wood properties
by using near-infrared laser radiation. For Prod J 50 1:81-86
Ulaby FT, Jedlicka RP (1984) Microwave dielectric properties of plant materials. IEEE Trans
Geosci Remote Sensing GE 22:406-414
Ulaby FT, Razni M, Dobson MC (1983) Effect of vegetation cover on microwave radiometric sen-
sitivity to soil moisture. IEEE Trans Geosci Remote Sensing GE 21:51-61
Ulaby FT, Sarabandi K, McDonald K, Whitt M, Dobson MC (1990) Michigan microwave canopy
scattering model. Int J Remote Sensing 11:1223-1253
Unger A, Planitzer J, Morgos A (1988) X-ray computer tomography and magnetic resonance
tomography for characterizing wet archaeological wood. Holztechnologie 29:249-250
USDA, Forest Product Laboratory, Forest Service (1972) Thermal properties. In: Wood handbook:
wood as an engineering material. USDA, Madison, WI, pp 324-326
Uyemura T (1960) Dielectric properties of wood as the indicator of the moisture. Bull Gov For
Exp Stn Tokyo 119:95-172
Vansteenkiste D (2000) Mise au point d'une methode rapide d'analyse quantitative de l'anatomie
du bois de cMne: analyse automatique des cliches radiographiques. In: Bobillier E, Jacob D
(eds) 6eme journee de la mesure - Electronique, Informatique, Automatique. INRA, Rennes,
8 pp
Varga A, Matusik A, Kiss E, Divos F (2000) Use of medical CT to qualify defects in wood. 12th
Symp Nondestructive Testing of Wood, Sopron, p 466
Vavilov V (1992) Thermal non destructive testing: short history and state of art. In: Balageas D,
Busse G, Carlomagno GM (eds) Proc Quantitative Infrared Thermograpfy (QIRT92). Editions
Europeennes Thermiques et Industrie, Paris, pp 179-194
Veres JS, Cofer GP, Johnson GA (1993) Magnetic resonance imaging of leaves. New Phytol 132:
769-774
Vermass HF (1973) Regression equations for determining the dielectric properties of wood. Holz-
forschung 27:132-136
Vincent L (1993) Morphological grayscale reconstruction in image analysis. Applications and effi-
cient algorithms. IEEE Trans Image Process 2:176-201
320 References
Vinegar HJ (1986) X-ray computed tomography and NMR imaging of rocks. J Petrol Technol
March:257-259
Volgyi F (2000) Microwave NDT of particleboards. 12th Symp Nondestructive Testing of Wood,
Sopron, pp 357-365
Vozzo JA, Halloin JM, Cooper TG, Potchen EJ (1996) Use of NMR spectroscopy and MRI for dis-
criminating Juglans nigra L. seeds. Sci TechnoI24:457-463
Wagner FG, Taylor FW (1985) Economic returns from internal log scanning. Fifth Nondestruc-
tive Testing of Wood Symp, Washington State University, Pullman, pp 267-280
Wagner FG, Taylor FW, Ladd DS, Mcmillin CW, Roder FL (1989a) Ultrafast CT scanning of an oak
log for internal defects. For Prod J 3911112:62-64
Wagner FG, Taylor F, Ladd D,McMillin C, Roder F (1989b) Ultrafast CT scanning oflogs for inter-
nal defects. 7th Symp Nondestructive Testing of Wood Symposium, Washington State Uni-
versity. Eng Publications, Pullman, pp 221-229
Wang PC, Chang SJ (1986) Nuclear magnetic resonance imaging of wood. Wood Fiber Sci 18:
308-314
Wang PC, Chang JS, Olson JR (1990). Scanning logs with an NMR scanner. 7th Symp Nonde-
structive Testing of Wood, Washington State University, Pullman, pp 209-219
Wang YQ, Kline RA (1994) Ray tracing in isotropic and anisotropic materials: application to tomo-
graphic image reconstruction. J Acoust Soc Am 95 2525-2532
Weatherwax RC, Stamm AJ (1947) The coefficients of thermal expansion of wood and wood prod-
ucts. Trans ASME May:421-432
Weibe S (1992) Untersuchungen zur Wundenwicklung und Wundbehandlung an Biiumen
unter besonderer Beruchsichtigung der Holzfeuchte. PhD thesis, Univ Miinchen, Miinchen,
Germany
Wert CA, Weller M, Caulfield D (1984) Dynamic loss properties of wood. J Appl Phys 56 9:
2453-2458
White MS, Ifju G, Johnson JA (1977) Methods for measuring resin penetration into wood. For
Prod J 27 7:52-54
Whittall KP, MacKay AL (1989) Quantitative interpretation ofNMR relaxation data. J Magn Reson
84:134-152
Wiberg P, Moren TJ (1999) Moisture flux determination in wood during drying above fiber
saturation point using CT scanning and digital image processing. Holz Roh Werkst 57:
137-144
Wiberg P (1996) CT-scanning during wood drying for nondestructive moisture content gradient
evaluation. Proc 10th Nondestructive Testing of Wood Symposium. Swiss Federal Institute of
Technology, Press Poly techniques et Universitaires Romandes, Lausanne, p 409
Wickman BE (1985) Comparison of the degree-day computer and a recording thermograph in a
forest environment. Reserch Note PNW-427. USDA, Forest Service, Madison, WI, 6 pp
Willis RL, Stone TS, Berthelot YH, Madigorski WM (1997) An experimental numerical technique
for evaluating the bulk and shear dynamic moduli of viscoelastic materials. J Acoust Soc Am
102:3549-3555
Willis RL, Wu L, Berthelot YH (2001) Determination of the complex Young and shear moduli of
viscoelastic materials. J Acoust Soc Am 109:611-621
Wilson DW, Charles JA (1981) Thermographic detection of adhesive-bond and interlaminar flaws
in composites. Exp Mech 21:276-280
Wilson MA, Godfrey 1M, Hanna JV, Quezada RA, Finnie KS (1993) The degradation of wood in
old Indian Ocean shipwrecks. Org Geochem 20 5:599-610
Winistorfer PM, Moschler WW Jr (1986) A direct scanning densitometer to measure density pro-
files in wood composite products. For Prod J 36 11112:82-86
Wisniewski M, Lindow SE, Ashworth EN (1997) Observations of ice nucleation and propagation
in plants using infrared video thermography. Plant Physiol113 2:327
Wissing ST, Wellinng J (1995) Nondestructive measurement of the density profile in wood based
panels by means of radio frequency. Proc Int Conf Progress in Forest Products Research,
Gottingen, p 165
References 321
Wolter B, Netzelmann U (1996) Nondestructive testing of wood using a single sided nuclear mag-
netic resonance. 10th Int Symp on Nondestructive Testing of Wood. Press Poly techniques et
Universitaires Romandes, Lausanne, p 409
Wu D (1992) Lockin thermography for multiplex photothermal nondestructive evaluation. In:
Balageas D, Busse G, Carlomagno GM (eds) Proc Quantitative Infrared Thermograpfy
(QIRT92). Editions Europeennes Thermiques et Industrie, Paris, France, pp 371-376
Wu D (1994) Lock in thermography for defect characterization in veneers. In: Balageas D, Busse
G, Carlomagno GM (eds) Proc Quantitative Infrared Thermograpfy (QIRT94). Editions
Europeennes Thermiques et Industrie, Paris, France, pp 298-301
Wu D, Busse G (1995) Remote inspection of wood with lock in thermography. Proc of 1995 Euro-
pean Plastic Laminates Forum "The leading Edge", Heidelberg, pp 27-29
Wu D, Busse G (1996) Remote inspection of wood with lock-in thermography. Tappi J 79
8:119-123
Wu D, Karpen W, Busse G (1992) Lockin thermography for multiplex photothermal nondestruc-
tive evaluation Proc Quantitative Infrared Thermography (QIRT 1992). Eurotherm Ser
27:37l-376
Wu D, Salerno A, Sembach, J, Maldague X, Rantala J, Busse G (1996) Wood-based products in-
spection by lockin thermography. Proc 3rd Eurowood Symp and 4th FESYP Technical Conf
Braunschweig, Germany, 5 pp
Wu D, Salerno A, Sembach J, Maldague X, Rantala J, Busse G (1997a) Lockin thermographic
inspection of wood particleboard. SPIE 3056:230-234
Wu D, Sembach J, Salerno A, Hora G, Busse G (1997b) Ensuring the quality of coated wood-based
products by means of lock-in thermography. (Qualitatssicherungg beschichteter Holzwerk-
stoffe mittels lockin-Thermographie). Holz -ZentralbI123:50-775
Wycoff W, Pickup S, Cutter B, Miller W, Wong TC (2000) The determination of the cell size in
wood by nuclear magnetic resonance diffusion techniques. Wood Fiber Sci 32:72-80
Xu Y, Okumura S, Noguchi M (1993) Thermographic detection of starved joints of wood. Mokuzai
Gakkaishi 39:544-549
Xu Y, Okumura S, Noguchi M (1994) Thermographic detection of starved joints of wood. 9th
Symp NDT of wood. Forest Product Soc, Madison, pp 209-217
Yamanaka K, Nagata Y, Koda T (1991) Enhancement of acoustic imaging of polymers by cooling.
Ultrasonics 29 March:159-165
Yokoyama M, Norimoto M (1996) Contour diagrams of dielectric loss for absolutely dried spruce
wood. Wood Res Kyoto Univ 83:37-39
Yokoyama M, Norimoto M (1997) Cole-Cole plots for dielectric properties of absolutely dried
wood. Tech Notes 33:71-82
Yokoyama M, Obataya E, Norimoto M (1999) Mechanical and dielectric relaxations of wood in
a low temperature range. II. Relaxations due to adsorbed water. Mokuzai Gakkaishi 45:95-
102
Yokoyama M, Kanayama Y, Furuta Y, Norimoto M (2000a) Mechanical and dielectric relaxations
of wood in a low temperature range. III. Application of search law to dielectric properties due
to adsorbed water. Mokuzai Gakkaishi 46: 173-180
Yokoyama M, Ohmae K, Kanayama K, Furuta Y, Norimoto M (2000b) Mechanical and dielectric
relaxations of wood in a low temperature range. IV. Dielectric properties of adsorbed water
at high moisture contents. Mokuzai Gakkaishi 46:523-530
Youngquist JA, Hamilton TE (1999) The next century of wood products utilization: a call for
reflection and innovation. Proc Int Conf on effective utilization of plantation timber, Chi -Tou,
Taiwan. For Prod Assoc ROC Bull 16:1-9
Zhu D, Beex AA, Conners R (1991) Stochastic field-based object recognition in computer vision.
SPIE 1569:174-181
Zielonka P, Dolowy K (1998) Microwave drying of spruce: moisture content, temperature and heat
energy distribution. For Prod J 48 6:77-80
Zimmermann MH (1983) Xylem structure and the ascent of sap. Springer, Berlin Heidelberg
New York, 143 pp
322 References
Zoughi R (1990) Microwave nondestructive testing: theories and applications. lnt Adv Nondestr
Test 15:225-288
Zoughi R (1996) Principles of and applications of microwave and millimeter wave NDE
methodologies: an overview. SPIE Proc 2944:46-74
Zoughi R, Ganchev S, Carriveau GW (1996) Overview of microwave NDE applied to thick com-
posites. Mater Sci Forum 210-213:69-76
Subject Index
Shape 4,31,37 C
Shrinkage 282 C-scan imaging 189
Structure 26, width 4,22,25,45-47,44, Calibration 51,169,174
45,100 Cambial zone 240,247,248
Water deficiency 287 Canopy 161
Antenna 163 Camera:
Archeological 69,259,273-276 CCD 156,157
Architecture of the scanner 28, 29 Video camera in visible light 75
First generation 30, 70 Infrared video camera 76,77,83,85,
Second generation 13,14,30 86,87,90,95,98,113,115,120,
Third generation 30, 70 122
Fourth generation 28,31,70 Capillary
Array of detector 13,19,29,32,36,55,70, Channels 247
109,186,196,208,238 Forces 234
Artifacts 149 Size 247
Beam hardening 22, 39, 40, 71, 72 Tube 242
Polychromaticity 40 Water 235
Artificial neural networks 25, 59 Carr-Purcell/Meiboom Gill sequence 226,
Aperture 35, 39 227,254,255,270,271
Approximation Born, Rytov 187 Cavity resonators 155, 178
Ash ions 33 Classification of NDT methods 4-9
A-scan imaging 189 Climatology 40, 69, 70
Atomic Cellulose 22,23, 34, 35, 125, 126, 142,
Number 14,70 144,220,232,256-259,268,279,
Arrangement 282 282
Reactor 283-285 Cellulose crystallinity 1,146,147,220
Attenuation of Ceramics 28,113,132,198
Microwave 128,133-140,144,150,161, Chemical
170,175-178 Cell wall 256-259
Neutron 282 Characterization 2,244, 279
Ultrasound 119,182,200,214 Constituents of wood 14,22,69,70,77,
X rays 19,20,23,24,28,38 137,146,147,175,282,287
Axes of elastic symmetry of wood, see Environment 217
longitudinal or radial or tangential Exchange 260
Imaging with liquid crystals 77, 122
B Shift 256, 268
B-scan imaging 189 Treatment 3,272,274
Bark 159,160,191,248,251,288,289 Christoffel equation 182, 183
Beam path 16, see ray Chip board 113
Bending loading 92, 112 Coalification of wood 275
Biologic Coding of tomograms
Degradation 49, 54, 69, 252 Color 14,43,44,90,99,106,108,119,120,
Specimen 49 121,182,183,185
Blackbody radiation 75 Gray scale 44
Bloch equations 218,252 Cole-Cole equation 145
Blows 192 Colimated beam 28, 44
Blister 192 Colimators 15
Bolzmann 217 Compensatory materials (paraffin, wax,
Bound water 12,135,172,173,224 granular sugar, acryl) 40
Branches 161 Compression loading 93,94,109-111
Bridge 28, 149 Compression wood 264, 265
Building Compton effect 71
Materials 294 Computation time for image reconstruction
Inspection 70 46
Subject Index 325
A thermal wavelength
e* complex dielectric constant
ro angular frequency of the intensity modulated laser heat source
o chemical shift
p density
't recovery time
(J screening constant
a thermal diffusivity
~<I> variation of phase of the radio frequency wave
e' and e" components of the complex dielectric constant (real and
imaginary)
~A variation of the amplitude of the radio-frequency wave
canopy equivalent dielectric constant
complex permittivity of dielectric material
free space perimittivity
Larmor frequency
phase constant of the air
wavelength of the electromagnetic wave in air
complex permittivity of water
attenuation coefficient of the ionizing radiation beam
thermal diffusion length
linear attenuation coefficient of X-rays for the x voxel in the cross-
sectional slice
!lw linear attenuation coefficient of X-rays in water
J.L' mass attenuation coefficient
J.Lc attenuation coefficient due to Compton effect
J.LP attenuation coefficient due to photo-electric effect
a cell lumen size (J.Lm) with thermal imaging
Bo static magnetic field
C specific heat of the solid wood
C image contrast
c speed of propagation of the radiation
C*(ro) complex capacitance
Cll> C22 , C33 ultrasonic stiffnesses on axes 1,2 and 3
d sample thickness (mm) with thermal imaging
DandL defect depth and length with thermal imaging
e thermal effusivity
336 List of Notations
b)
24
o
36--- -- ----
42
48
54
60
66
72
78
84
90
96 24
30
6
-----------
42
48
>466
60 72
790
84 96
Fig.2.17a,b. Tomogram on a color scale of a transverse section of a healthy (a) and of a decayed
(b) Norway spruce tree. (Habermehl and Ridder 1996, with permission)
a)
M..
... • .• ..W
...
--
• .W ~ ..
.. .M.. • .•
Fig.2.23a,b. Tomograms of Quercus petraea and Quercus cerris (Tognetti et al. 1996). a Quercus
--
......
cerris; b Quercus petraea. N, E, S, Ware the cardinal points; 1,2,3,4 are the position of measured
points The density scale in CT numbers is represented by different colors (i.e., CT = 30 corre-
sponds to dry condition and CT = 96 corresponds to high moisture content). (with permission)
338 Color Plates
24.00
30.00
36.00
-
42.00
48.00
54.00
60.00
66.00
72.00
-
78.00
84.00
90.00
96.00
24.00
30.00
36.00
42.00
48.00 --
54.00
60.00
66.00
72.00
78.00
84.00
Fig.2.25a-d. The development of sapwood in Scots pine induced by ammonia pollution, com-
90.00
pared with a normal tree at 100cm stem height. Trees are located in a forest from the district
96 . 00
Torgelow, Germany (Katzel et al. 1997, with permission). A Tomogram for tree no. 1, with 624-
cm' cross section located at site 1,200 m from the farm (maximum pollution zone). B Tree no. 8,
with 745-cm' cross section located at site 2, 280m from the farm. C Tree no. 2 at site 3, with 513-
cm' cross section located 2900m from the farm (minimum pollution zone). D Tree no. 8 at site 3
located 2900m from the farm (minimum pollution zone). The density scale in CT numbers is
represented by different colors, CT = 30 corresponds to dry conditions and CT = 96 corresponds
to a high moisture content. The tomograms were taken in September 1993 and 1994
a b c
36~
36 36_ - 85I)oil5O
1 -490-S40
30 _~ .• 540-590
30 _~ 7~
.. &50-?50
30_ . 1~"00 .300-440
24 i·~~
24 .,QOO-'QM 24i ~= 0_
i • 950-1000
~ 0350-450
18 18 18 5
5
12=
:2! ~2!
6 !
0 0 0
0 10 20 30 40 50 60 70 eo 90 100 0 10 20 30 40 50 60 70 eo 90 100 0 10 20 30 40 50 60 70 80 90 100
d Board width (nvn) e Board width (nvn) r Board width (mm)
: 36~ 36~
36 8().ll1O ~ . '4».1 20
I ·, 30 ~ . 15-18
30_ :~:~: 30 - •. 8().100
eo-eo - . '2·\$
: 24i! ~::: 24 ~ •• ~12
6·9
241 ~ 0 0·20 o 3 ·s
18 . 18 . 18 .
S :;
Fig.2.32a-i. Dynamics of wood drying (Pang and Wiberg 1998). Experimental tomograms (with permission). a Scanning of the board before drying -
corresponding to wet saturated wood. b Scanning of the board before drying - corresponding to wet wood after 9.6 hours of drying, and to the early stage
of the drying process. c Scanning of the board before drying - corresponding to wet wood after 3004 h of drying and to the late stage of the drying process.
Predicted images for density distribution. d Wet wood density before drying. e Wood density after 9.6 h of drying. f Wood density after 30.4 h of drying.
<J.>
Predicted images for moisture content distribution. g Model of predicted wet wood density before drying. h Model of predicted wet wood density after <J.>
-c
9.6h of drying. i Model of predicted wood density after 30Ah of drying
340 Color Plates
(unit: mm)
Fig. 3.17. Thermographic measurements of the slope of grain. Finite element model. The heat
flux is represented by the vertical arrows in the central region. (Naito et al. 2000; courtesy of the
Japan Wood Research Society).
Fig. 3.18. Temperature distribution as a function of heating and slope of grain deduced with the
finite element method time obtained with the finite element method. (Naito et al. 2000; courtesy
of the Japan Wood Research Society)
342 Color Plates
0
Density (g/cm 3)
2
1.3000 - 1.4000
1.2000 - 1.3000
4 1.1 000 - 1.2000
-
1.0000 - 1.1000
6 0.9000 - 1.0000
0.8000 - 0.9000
8 0.7000 - 0.8000
0.6000 - 0.7000
10 0.5000 - 0.6000
0.4000 - 0.5000
12 0.3000 - 0.4000
0.2000 - 0.3000
14 0.1 000 - 0.2000
0-0.1000
0 2 4 6 8 10 12 14
a) x (mm)
3.0
Xylem vessels
2.5
Transmission
2.0 _ 567 - 600
535 - 567
1.5 502 - 535
470 - 502
1.0 437 - 470
405 - 437
0.5
372 - 405
_ 340-372
0 _ 308 - 340
0 2 3 4 5
b) _ 275 -308
x (mm)
Fig.3.20a,b. Density mapping of a transverse section of two species: a beech and b balsa.
The size of the specimen was 14 x 14 x 1.7mm. (Koch et a1.l998; courtesy of Wood Science
Technology)
Color Plates 343
Spec,"*, I*IaId
a)
diameter ~
(')
o
0"
....
:s
280 ~
'"
a) 300
150
g100
~ 50
5 • 'ir¥ .
4~ ....,...
~;j~: .'.'
~<""'''''.' ;.'.~
'..,'.!~,; ~
o
o 5 10 15 20 ;~~:~?? ~~ ~.,.>:. )'
b) OefIecUon(rrm) c)
Fig.3.32a-c. Infrared imaging of specimen under static bending loading. a Specimen with a central hole under static bending. b Stress strain
curve. c Temperature distribution as a function of loading and strain, corresponding to points 1, 2, 3, and 4. (Naito et a1.1998; courtesy of Wood
Research Institute Kyoto University)
Color Plates 347
Dome cr
I
\
~,
~
.'
,,"",
.'
I .
, ,
,
~, (mm)
10 .' .,
I
.,
.'
,
~
.'
, ., ,
.'
I
,
.'
I
t, I. ,
• \ .,
,- ,. , ,.,
'.' .' '.' .' '\, \' \ j,
1:' ,:1
:
':1
:
I:'
::
I .. ,
::
4 .'
~ \ \,;
~. .
::
a)
2
b) '- ... .\ _.6." II.
,
0 mm 2 mm
Fig.3.38a, b. Detection of artificially induced
defects in chipboard through the surface veneer
sheets of various thicknesses. a Geometry of
the specimen. b Thermal image. (Wu and Busse
1996, with permission)
a) b)
Fig.3.39a, b. Detection of knots present in solid wood covered with veneer sheets. a Geometry
of the specimen. b Thermal image. (Wu and Busse 1995, with permission)
348 Color Plates
Fig.3.41a-c. Infrared images of splice joints under shear loading. a Before loading. b Shear stress
before failure. Color hue corresponds to 0.2°C. (Luong 1996, with permission)
Color Plates 349
Fig.4.31a-d. Iso-surface on a topographic slice. a Ipi The amplitude; b ythe argument (y, arg p).
c e The ellipticity angle which gives the phase difference between two wave components and
describes the degree of elliptical polarization. d r The angle between the two components of
the wave, corresponding to the tilt of the linear polarization. (Kaestner and Baath 2000, with
permission)
350 Color Plates
Fig.4.33a,b. Images of knots reconstructed with a the polarization ratio; b the phase difference
between the two components of the electric field, parallel and perpendicular to the fibers.
(Kaestner and Baath 2000, with permission)
Color Plates 351
Polan ation
along e
(transverse wave)
Polarisation
along r
(longitudinal wave)
Polari ation
along '"
(transverse wave)
b)
Oak
0.5
-0.5
-1
1.00
0.50
[mm]
Fig. 5.5. High resolution images of the transverse section of a tree (Platanus acerifolia)
(Martinis 2002, with permission)
km/s
600 2.00
1.80
500
1.60
400 1.40
E 1.20
§. 300
1.00
200 0.80
0.60 Fig. 5.17. Tomographic
100 images of the transverse
0.40 section of the trunk in Fig.
0.20 5.16 obtained with 120
100 200 300 400 500 600 independent velocity
measurements. (Comino et
[mm] al. 2000, with permission)
Color Plates 353
a)
600
1.80
SOD
1.60
400 1.40
E 1.20
~ 300
1.00
200
0.80
100 0.60
0.40
100 200 300 400 500 600
b) [mm]
2
1.8
I
1.S
1.4
1.2
~
~
0
J 0 .8
0.8
0.4
sample labels
Fig. 5.19. Measured ultrasonic velocities and calculated velocities for three orthotropic direc-
tions, L, Rand T on selected specimens shown in Fig. 5.1S. (Socco et al. 2000, with
permission)
354 Color Plates
a b c