CN103398957A - Hyperspectrum and laser radar-based method for extracting vertical distribution of leaf area - Google Patents

Abstract

The invention discloses a hyperspectrum and laser radar-based method for extracting vertical distribution of leaf area. The hyperspectrum and laser radar-based method mainly comprises the steps of 1, classifying point cloud data of an airborne laser radar, and extracting vegetation structure parameters; 2, acquiring distribution of region leaf area indexes by virtue of a geometrical optical model on the basis of hyperspectrum data and the extracted vegetation structure parameters; 3, calculating the percentage of vegetation laser points above the ground on each height layer, thus obtaining a height profile of a corresponding vegetation canopy; and 4, on the basis of extracting the vegetation leaf area indexes and the height profile of the canopy, distributing the vegetation leaf area indexes according to the height profile of the canopy to obtain a vegetation leaf area index of each layer and vegetation leaf area indexes accumulated along with the height. According to the hyperspectrum and laser radar-based method, horizontal information of the hyperspectrum data and response of the laser radar to vegetation height information are comprehensively utilized, the vertical distribution of the region leaf area indexes is extracted, and more accurate parameter input is provided for a physical model-based vegetation radiation transmission model.

Description

Method based on high spectrum and laser radar extraction leaf area vertical distribution

Technical field

The invention belongs to ecological vegetation calculation of parameter assessment technology field, relate in particular to the extracting method of leaf area index vertical distribution.

Background technology

Leaf area index (LAI) is required important vegetation parameters of some relevant ecological processes, its vertical distribution affects photosynthetically active radiation, the photosynthesis of vegetation and evapotranspires, and it is also one of important criterion of Forest Carbon revenue and expenditure, therefore to its precise evaluation for evapotranspiring and Net primary productivity is assessed all important roles.

For the clear vertical stratification of describing the vegetation canopy, the research of leaf altitude profile or canopy height section (Canopy Height Profile, CHP) is seemed to very important, cause thus the interest of people to the research of leaf area index vertical distribution.Some researchers have directly studied the leaf area index section, and other researchists are reflected the leaf area vertical distribution by research leaf area density or tree and grass coverage density vertical distribution.Because leaf area index obtains by cumulative leaf area density numerical evaluation, so the vertical distribution of leaf area also can be described by the leaf area density function of each level height layer.For the assessment of the vertical distribution of leaf area, mainly contained leaf area collection and model two kinds of modes of assessment indirectly in the past, directly acquisition method usually will relate to and destroy canopy and gather problem consuming time, and be also inconvenient in the enough samplings of collection of large tracts of land test block, this method also is not suitable for the long term monitoring of leaf area on room and time is dynamic; The measuring accuracy of rear a kind of mode usually can be subject to the space distribution of leaf and the restriction of illumination condition.In recent years, the use of laser radar in forest makes people, to the vertical stratification of canopy, more understanding arranged.

Summary of the invention

The present invention is directed to the deficiencies in the prior art, a kind of method based on high spectrum and laser radar extraction leaf area index vertical distribution is provided, the horizontal information of the method comprehensive utilization high-spectral data and the reflection that the laser class reaches the vegetation elevation information, the vertical distribution of the regional leaf area index of extraction

For realizing above technical purpose, the present invention will take following technical scheme: a kind of method based on high spectrum and laser radar extraction leaf area index vertical distribution comprises the following steps:

Step 1: airborne laser radar point cloud data is classified according to ground point and vegetation point, extract the vegetation structure parameter in corresponding high spectrum pixel;

Step 2: based on the vegetation structure parameter of high-spectral data and extraction, and adopt geometric optical model to obtain regional leaf area index distribution;

Step 3: the vegetation laser spots more than ground is calculated to its number percent on each height layer, obtained corresponding vegetation canopy height section;

Step 4: on the basis of extracting vegetation leaf area index and canopy height section, distribute the vegetation leaf area index according to the canopy height section, obtain the vegetation leaf area index of every one deck and with the vegetation leaf area index of highly accumulating.

As preferably, vegetation structure parameter shown in step 1 comprises the height of tree, hat width and clear bole height.

As preferably, the preparation method that the described regional leaf area index of step 2 distributes is specially:

At first, the Li-Strahler geometric optical model is reduced to two component Models shown in formula (1) to the earth surface reflection signal description on remote sensing image:

S＝K _gG+K _cC （1）

Wherein, S is the earth surface reflection signal on remote sensing image, and G, C are respectively the reflected signals of illumination earth's surface and illumination tree crown, by the pure pixel spectrum that obtains, is obtained; K _c, K _gIt is respectively corresponding area percentage;

Then, the structural parameters in conjunction with geometric optical model and laser radar extraction, extract leaf area index by formula (2), (3) and (4),

$\cos t=\frac{h|{\tan \mathrm{\θ}}_i-{\tan \mathrm{\θ}}_v\cos \mathrm{\φ}|}{r({\sec \mathrm{\θ}}_i+{\sec \mathrm{\θ}}_v)}---\left(2\right)$

$M=\frac{-\ln \left(K_g\right)}{({\sec \mathrm{\θ}}_i+{\sec \mathrm{\θ}}_v)(\mathrm{\π}-t+\cos t\sin t)}---\left(3\right)$

LAI＝2πM （4）

Wherein, θ _i, θ _vBe respectively the zenith angle of the sun and satellite;

It is the relative bearing between the sun and satellite; H and r are respectively mean height and the even crown diameter radiuses of vegetation; T is angle, and scope is [0, pi/2], and M is that the average canopy in sample ground covers size, and LAI is leaf area index.

As preferably, the vegetation coverage described in step 3 on each height layer represents with the number percent of vegetation laser spots, as shown in Equation (5):

$P_v\left(h\right)=\frac{{\mathrm{Num}}_v\left(h\right)}{{\mathrm{Num}}_t}---\left(5\right)$

Wherein, P _v(h) be the vegetation coverage on each height layer, Num _v(h) the above all vegetation laser spots numbers of expression height h, Num _tRepresent all laser spots numbers;

By formula (6), obtain accumulating the canopy height section:

CHP(h)＝-ln(1-P _v(h)) （6）

Wherein, CHP (h) is the canopy cumulative percentage at h place;

By formula (7), obtain the canopy height section:

$\mathrm{relCHP}\left(h\right)=\frac{\mathrm{CHP}(h-1)-\mathrm{CHP}\left(h\right)}{\mathrm{CHP}\left(0\right)}---\left(7\right)$

Wherein, rel CHP (h) is the canopy height section at h place, the canopy cumulative percentage of CHP (0) the expression bottom.

According to above technical scheme, compared with prior art, the present invention has advantages of following: adopt high-spectral data and laser radar to extract regional leaf area index, then the vegetation canopy height section that extracts in conjunction with laser radar has obtained the vertical distribution of study area leaf area index, and more accurate parameter input can be provided for the vegetation radiative transfer model of Physical modeling based.

The accompanying drawing explanation

Fig. 1 is the method flow schematic diagram based on high spectrum and laser radar extraction leaf area index vertical distribution shown in the present.

Embodiment

Accompanying drawing discloses the structural representation of preferred embodiment involved in the present invention without limitation; Below with reference to accompanying drawing, explain technical scheme of the present invention.

The airborne laser radar that remotely-sensed data is obtained middle use is the Litemapper-5600 instrument that German IGI company produces, and in study area, on the about 700-800m height in ground, has obtained laser radar data, and packing density is 0.36-1.6/m ²Individual point.High-spectral data is the Hyperion data, the leaf area index data of ground data for adopting LAI2000 to measure.

As shown in Figure 1, the specific implementation step is:

Step 1, carry out the classification of ground point and vegetation point to airborne laser radar point cloud data, extract the vegetation structure parameter in corresponding high spectrum pixel, comprises the parameters such as the height of tree, hat width, clear bole height.

Step 2, adopt geometric optical model to obtain regional leaf area index based on the vegetation structure parameter of high-spectral data and extraction and distribute; The Li-Strahler geometric optical model is described and is reduced to two component Models the remote sensing signal of mixed pixel:

S＝K _gG+K _cC （1）

G wherein, C be respectively illumination earth's surface, illumination tree crown, reflected signal; K _c, K _gIt is respectively corresponding area percentage.G, the C component is obtained by the pure pixel spectrum that obtains.Structural parameters in conjunction with geometric optical model and laser radar extraction extract leaf area index, and formula is:

$\cos t=\frac{h|{\tan \mathrm{\θ}}_i-{\tan \mathrm{\θ}}_v\cos \mathrm{\φ}|}{r({\sec \mathrm{\θ}}_i+{\sec \mathrm{\θ}}_v)}---\left(2\right)$

$M=\frac{-\ln \left(K_g\right)}{({\sec \mathrm{\θ}}_i+{\sec \mathrm{\θ}}_v)(\mathrm{\π}-t+\cos t\sin t)}---\left(3\right)$

LAI＝2πM （4）

θ wherein _i, θ _vBe respectively the zenith angle of the sun and satellite,

Be the relative bearing between the sun and satellite, h and r are respectively average height, the even crown diameter radiuses of vegetation.

Step 3, calculate its number percent and obtain corresponding vegetation canopy height section on each height layer to the vegetation laser spots more than ground; Highdensity airborne LiDAR point cloud can be used for representing the covering of vegetation, and therefore the percentage for the available vegetation laser spots of vegetation covering on each height recently represents, as shown in the formula:

$P_v\left(h\right)=\frac{{\mathrm{Num}}_v\left(h\right)}{{\mathrm{Num}}_t}---\left(5\right)$

Num wherein _v(h) the above all vegetation laser spots numbers of expression height h, Num _tRepresent all laser spots numbers.

Changing into accumulation canopy height section is:

CHP(h)＝-ln(1-P _v(h)) （6）

Being converted into the canopy height section is:

$\mathrm{relCHP}\left(h\right)=\frac{\mathrm{CHP}(h-1)-\mathrm{CHP}\left(h\right)}{\mathrm{CHP}\left(0\right)}---\left(7\right)$

Wherein CHP (0) expression is the canopy cumulative percentage of the bottom, is also maximum canopy number percent accumulated value.

Step 4, on the basis of extracting vegetation leaf area index and canopy height section, distribute the vegetation leaf area index according to the canopy height section, obtains the vegetation leaf area index of every one deck and with the vegetation leaf area index of highly accumulating.

Claims (4)

1. method of extracting the leaf area index vertical distribution based on high spectrum and laser radar comprises the following steps:

Step 1: airborne laser radar point cloud data is classified according to ground point and vegetation point, extract the vegetation structure parameter in corresponding high spectrum pixel;

Step 2: based on the vegetation structure parameter of high-spectral data and extraction, and adopt geometric optical model to obtain regional leaf area index distribution;

Step 3: the vegetation laser spots more than ground is calculated to its number percent on each height layer, obtained corresponding vegetation canopy height section;

Step 4: on the basis of extracting vegetation leaf area index and canopy height section, distribute the vegetation leaf area index according to the canopy height section, obtain the vegetation leaf area index of every one deck and with the vegetation leaf area index of highly accumulating.

2. according to claim 1 based on the method for high spectrum and laser radar extraction leaf area index vertical distribution, it is characterized in that: vegetation structure parameter shown in step 1 comprises the height of tree, hat width and clear bole height.

3. according to claim 2 based on the method for high spectrum and laser radar extraction leaf area index vertical distribution, it is characterized in that: the preparation method that the described regional leaf area index of step 2 distributes is specially:

At first, the Li-Strahler geometric optical model is reduced to two component Models shown in formula (1) to the earth surface reflection signal description on remote sensing image:

S＝K _gG+K _cC （1）

Wherein, S is the earth surface reflection signal on remote sensing image, and G, C are respectively the reflected signals of illumination earth's surface and illumination tree crown, by the pure pixel spectrum that obtains, is obtained; K _c, K _gIt is respectively corresponding area percentage;

Then, the structural parameters in conjunction with geometric optical model and laser radar extraction, extract leaf area index by formula (2), (3) and (4),

$\cos t=\frac{h|{\tan \mathrm{\θ}}_i-{\tan \mathrm{\θ}}_v\cos \mathrm{\φ}|}{r({\sec \mathrm{\θ}}_i+{\sec \mathrm{\θ}}_v)}---\left(2\right)$

$M=\frac{-\ln \left(K_g\right)}{({\sec \mathrm{\θ}}_i+{\sec \mathrm{\θ}}_v)(\mathrm{\π}-t+\cos t\sin t)}---\left(3\right)$

LAI＝2πM （4）

Wherein, θ _i, θ _vBe respectively the zenith angle of the sun and satellite;

It is the relative bearing between the sun and satellite; H and r are respectively mean height and the even crown diameter radiuses of vegetation; T is angle, and scope is [0, pi/2], and M is that the average canopy in sample ground covers size, and LAI is leaf area index.

4. based on high spectrum and laser radar, extract the method for leaf area index vertical distribution shown according to claim 3, it is characterized in that: the vegetation coverage described in step 3 on each height layer represents with the number percent of vegetation laser spots, as shown in Equation (5):

$P_v\left(h\right)=\frac{{\mathrm{Num}}_v\left(h\right)}{{\mathrm{Num}}_t}---\left(5\right)$

Wherein, P _v(h) be the vegetation coverage on each height layer, Num _v(h) the above all vegetation laser spots numbers of expression height h, Num _tRepresent all laser spots numbers;

By formula (6), obtain accumulating the canopy height section:

CHP(h)＝-ln(1-P _v(h)) （6）

Wherein, CHP (h) is the canopy cumulative percentage at h place;

By formula (7), obtain the canopy height section:

$\mathrm{relCHP}\left(h\right)=\frac{\mathrm{CHP}(h-1)-\mathrm{CHP}\left(h\right)}{\mathrm{CHP}\left(0\right)}---\left(7\right)$

Wherein, rel CHP (h) is the canopy height section at h place, the canopy cumulative percentage of CHP (0) the expression bottom.