Object-based image analysis

Illumination invariance remains the most researched, yet the most challenging aspect of automatic face recognition. In this paper we investigate the discriminative power of colour-based invariants in the presence of large illumination changes between training and test data, when appearance changes due to cast shadows and non-Lambertian effects are significant. Specifically, there are three main contributions: (i) we employ a more sophisticated photometric model of the camera and show how its parameters can be estimated, (ii) we derive several novel colour-based face invariants, and (iii) on a large database of video sequences we examine and evaluate the largest number of colour-based representations in the literature. Our results suggest that colour invariants do have a substantial discriminative power which may increase the robustness and accuracy of recognition from low resolution images.

In this paper, we describe the application of Object Based Image Analysis for the knowledge-based, automated mineral
classification from petrographic micrographs. Digital color images, acquired by an optical, petrographic microscope at
different stage rotations and with parallel and crossed polarization filters, are the input to a sequence of different image
segmentation and fuzzy classification steps. As compared with traditional, pixel-based algorithms, Object Based Image
Analysis incorporates not only the spectral characteristics of various mineral phases, but also their topological and genetic
properties, resulting in superior image classification results. By means of rule sets, image analysis can be flexibly adapted
to different rock types.

Data mining and knowledge discovery in databases have been attracting a significant amount of research, industry, and media attention of late. There is an urgent need for a new generation of computational theories and tools to assist researchers in extracting useful information from the rapidly growing volumes of digital data

Automatic attendance system is one of the significant issues of today's research. Among other methods, human face recognition is highly used technique for attendance automation. Many systems have been proposed in literature using face recognition. Most of the systems are using fixed camera and desktop computers. We propose a system using mobile phones where an image is captured of group of peoples and face detection is done automatically. While considering computational and storage power of mobile devices, extracted local binary features for detected faces are then transferred to server machine using firebase database. Matching is done on server side, if face recognized than attendance is marked and feedback is sent back to client side. Experiments show effectiveness of proposed techniques with 95% correct recognition rate.

The belief that practitioner knowledge is tacit and cannot be communicated through the traditional methods of academic writing and journal publication is a problem for the growing number of fashion researchers utilising practice-led methodologies. The main reason for this is that, until recently, the visibility of a discipline has been predominantly defined by academic publications in the field. One solution has been for researchers to adopt research methodology that utilises reflective practice to draw out practitioner knowledge and attempt to convert this knowledge to an explicit form. Alternatively, non-traditional research outputs are now considered equivalent to published research within areas such as art and design. However, a critical mapping of fashion research reveals that practice-led research in fashion is not yet adequately represented by either form of research outcome. This thesis proposes that continuing to focus on the differences between practice-led research and other fashion research, a dychotomic view of practitioner knowledge as either tacit or explicit, and an upward trend in process-driven methodology, are limiting the ability for practitioners to contribute to the continuing development of fashion as a discipline. The aim of this thesis is to develop a progressive approach to practice-led research methodology to enable practitioner knowledge of fashion to become more accessible, transferable and relevant to the emerging methodology of fashion as a material culture that connects two different ideas of the fashion system as either material or immaterial.

A quantitative and qualitative mapping of fashion research, based on journal publications, non-traditional research outputs and an analysis of practice-led research exegeses, determines the current state of fashion as an emerging discipline. The mapping reveals that dominant methodologies for existing practice-led research are design methodologies. The results from this mapping inform an exploration of the potential for object or artefacts to encapsulate fashion knowledge through the method of object analysis as an object-based methodology used within other areas of fashion research. The results of this analysis enable the development of two models for practice-led fashion research. The first is a model of practitioner knowledge and theorises tacit knowledge as a hybrid space existing as a relationship between the different types of knowledge that exist in fashion practice. The second is a model of practitioner fashion research that incorporates existing design methodology but also responds to emerging fashion theory and methodology. Both models extend current understanding of fashion practice and contribute to developing the most effective strategy for communicating practitioner knowledge of fashion within the university environment. 

EXTRACTİON OF DETAİLS İN THE CİTY CENTER BENEFİTİNG FROM REMOTE SENSİNG İMAGES USİNG BY OBJECT-BASED CLASSİFİCATİON METHOD

Nowadays, reference maps should be updated constantly for the reasons of preventing on the illegal, unplanned cities and improve the decision-making process. For this reason, in recent years, there has been a significant improvement in classification methods and satellite images for extraction data quickly and accurately. Due to the generation of centimeter-accurate images, new image classification methods have been developed that give high accuracy results for these new high accuracy images. One of those image classification techniques is object based classification methods that extract data with high accuracy from high resolution satellite images. In this thesis accurate and quickly extraction of details (building, roads, vegetation etc.) from urban areas with object based classification methods is aimed by using ortophotos and World View-3 high resolution satellite image and both have same resolution. In addition that, vector data extraction from both images is examined considering building details for the availability in Geographical Information Systems (GIS) based high scale studies. Therefore, generalization of building details in vector format which are extracted with object based classification method are compared statistically with current base map and digitized data from high resolution satellite images. Result of this study, data extraction results from object based classification method in world view 3 has higher accuracy than ortophotos. Besides, 20 buildings randomly selected from world view 3 classification results and their corner coordinates are compared with coordinates from digitized data and base map statistically. Results show that all values are in confidence interval. It reveals that Automatically Extraction of building details providing positive contribution in case of higher spatial resolution and higher value of spectral reflectance in satellite images.

Aerial reconnaissance continues to be a vital tool for landscape-oriented archaeological research. Although a variety of remote sensing platforms operate within the earth's atmosphere, the majority of aerial archaeological information is still derived from oblique photographs collected during observer-directed reconnaissance flights, a prospection approach which has dominated archaeological aerial survey for the past century. The resulting highly biased imagery is generally catalogued in sub-optimal (spatial) databases, if at all, after which a small selection of images is orthorectified and interpreted. For decades, this has been the standard approach. Although many innovations, including digital cameras, inertial units, photogrammetry and computer vision algorithms, geographic(al) information systems and computing power have emerged, their potential has not yet been fully exploited in order to re-invent and highly optimise this crucial branch of landscape archaeology. The authors argue that a fundamental change is needed to transform the way aerial archaeologists approach data acquisition and image processing. By addressing the very core concepts of geographically biased aerial archaeological photographs and proposing new imaging technologies, data handling methods and processing procedures, this paper gives a personal opinion on how the methodological components of aerial archaeology, and specifically aerial archaeological photography, should evolve during the next decade if developing a more reliable record of our past is to be our central aim. In this paper, a possible practical solution is illustrated by outlining a turnkey aerial prospection system for total coverage survey together with a semi-automated back-end pipeline that takes care of photograph correction and image enhancement as well as the management and interpretative mapping of the resulting data products. In this way, the proposed system addresses one of many bias issues in archaeological research: the bias we impart to the visual record as a result of selective coverage. While the total coverage approach outlined here may not altogether eliminate survey bias, it can vastly increase the amount of useful information captured during a single reconnaissance flight while mitigating the discriminating effects of observer-based, on-the-fly target selection. Furthermore, the information contained in this paper should make it clear that with current technology it is feasible to do so. This can radically alter the basis for aerial prospection and move landscape archaeology forward, beyond the inherently biased patterns that are currently created by airborne archaeological prospection.

The increasing amounts and variety within remote sensing data have triggered researchers to develop semi-automated methods for detecting landscape features. For archaeology it has been recognised that it is difficult to capture the unique characteristics of archaeological remains detected in the landscape. This has led to a discussion of the potential from which was concluded that variable features could not be captured. Nonetheless within geosciences recent success of pattern recognition allowed for the supervised classification of highly irregular landslides using cognitive reasoning and machine learning. These concepts form the basis of artificial intelligence (AI) which has yet to be explored for remote sensing in archaeology. The success in geoscience is largely provided by the geographical object based image analysis software, eCognition, which has been explored by archaeologists to some extent but was not applied in great detail. With the presentation of a pilot study for barrow recognition around Avebury, United Kingdom, it has been shown that the potential of this software is far reaching and requires a re-evaluation within the archaeological discipline for remote sensing. From three basic analysis based on cognitive ruleset, template matching and machine learning it was concluded that the first is simple to implement but requires a lot of manual refinement and with the second, simple and consistent features could be detected. Even though the machine learning algorithm did not yield the best results it is a known fact that with adequate amounts of training data, this is the best method to detect variable features as it writes the rules itself based on similarities between training samples. As it was shown that AI concepts are able to detect unique characteristics of archaeological features and combine multiple remote sensing layers to do so it was found that the future of semi-automated pattern recognition in archaeological remote sensing could greatly aid large scale mapping programmes. Additionally, these methods could be applied for change detection to monitor archaeological sites for improved heritage management and have the potential to be widely implemented in an all-encompassing online web application for exchange of rulesets and training data for generally known features.

Recent advancement in sensor technology allows very high spatial resolution alongwith multiple spectral bands. There are many studies, which highlight that Object Based Image Analysis(OBIA) is more accurate than pixel-based classification for high resolution(< 2m) imagery. Image segmentation is a crucial step for OBIA and it is a very formidable task to estimate optimal parameters for segmentation as it does not have any unique solution. In this paper, we have studied different segmentation algorithms (both mono-scale and multi-scale) for different terrain categories and showed how the segmented output depends upon various parameters. Later, we have introduced a novel method to estimate optimal segmentation parameters. The main objectives of this study are to highlight the effectiveness of presently available segmentation techniques on very high resolution satellite data and to automate segmentation process. Pre-estimation of segmentation parameter is more practical and efficient in OBIA. Assessment of segmentation algorithms and estimation of segmentation parameters are examined based on the very high resolution multi-spectral WorldView-3(0.3m, PAN sharpened) data.

One persistent archaeological challenge is the generation of systematic documentation for the extant archaeological record at the scale of landscapes. Often our information for landscapes is the result of haphazard and patchy surveys that stem from opportunistic and historic efforts. Consequently, overall knowledge of some regions is the product of ad hoc survey area delineation, degree of accessibility, effective ground visibility, and the fraction of areas that have survived destruction from development. These factors subsequently contribute unknown biases to our understanding of chronology, settlements patterns, interaction, and exchange. Aerial remote sensing offers one potential solution for improving our knowledge of landscapes. With sensors that include LiDAR, remote sensing can identify archaeological features that are otherwise obscured by vegetation. Object-based image analyses (OBIA) of remote sensing data hold particular promise to facilitate regional analyses thorough the automation of archaeological feature recognition. Here, we explore four OBIA algorithms for artificial mound feature detection using LiDAR from Beaufort County, South Carolina: multiresolution segmentation, inverse depression analysis, template matching, and a newly designed algorithm that combines elements of segmentation and template matching. While no single algorithm proved to be consistently superior to the others, a combination of methods is shown to be the most effective for detecting archaeological features.

Object based image analysis (OBIA) is a method of assessing remote sensing data that uses morphometric and spectral parameters simultaneously to identify features in remote sensing imagery. Over the past 10-15 years, OBIA methods have been introduced to detect archaeological features. Improvements in accuracy have been attained by using a greater number of morphometric variables and multiple scales of analysis. This article highlights the developments that have occurred in the application of OBIA within archaeology and argues that OBIA is both a useful and necessary tool for archaeological research. Additionally, I discuss future research paths using this method. Some of the suggestions put forth here include: pushing for multifaceted research designs utilizing OBIA and manual interpretation, using OBIA methods for directly studying landscape settlement patterns, and increasing data sharing of methods between researchers.

The data obtained from remote sensing satellites furnish information about the land at varying resolutions
and has been widely used for change detection studies. There exist a huge number of change detection
methodologies and techniques with the continual emergence of new ones. This paper provides a review of
pixel based and object-based change detection techniques in conjunction with the comparison of their
merits and limitations. The advent of very-high-resolution remotely sensed images, exponentially increased
image data volume and multiple sensors demand the potential use of data mining techniques in tandem
with object-based methods for change detection.

Sejalan dengan penelitian yang sedang dilakukan, saya ingin berbagi  pengetahuan dalam bidang Object Based Image Analysis. Tutorial ini berisi step by step menggunakan software ecognition 8.0 yang akan menjelaskan dasar penggunaan OBIA.

Well this is indonesian version, but if you interest i can make the english one. Hope you enjoy

"This research paper compares the result of Object based and Pixel based classification techniques for glacier change detection on Landsat Thematic mapper (TM) and Enhanced Thematic Mapper (ETM+) imageries. The objective of this study is to see which classification method performs better for change detection in mountainous regions. Northern face of Himalayan region, The study area is undergoing climate change in the form of rapid melting of glacial ice mass, expansion of the existing lakes and creation of new lakes. This results in Glacial Lake Outburst Flood (GLOF) and breach or outburst from ice
and ‘moraine dams’ causing devastating floods
downstream. The global warming phenomenon worldwide has resulted in a significant decrease in glacial cover.
Glacier’s change monitoring
and permafrost-related hazards have long been studied using remote sensing data and techniques to assess the damage. The world is facing a serious problem of handling the climate change issue and its effects on humans as well as on natural resources. Glaciers are considered as one of the best indicators of climate change [1]. Landsat TM/ETM+ images were used for glacier
change monitoring of Turkey’s mountains project, Mount
Suphan. The results show that about ¾ of total area of suphan glacier has been lost in 23 years. Traditional image classification methods use only the spectral information at pixel level without considering the shape of underlying objects [2]. However, object-based image classification process uses spectral and spatial dimensions (shape of feature) in order to perform classification. In this study, multi temporal Landsat TM and ETM+ image from 1990 to 2010 have been used. Initially, the traditional pixel-based classification was performed on Landsat thematic layers and layers developed from indices like NDVI and NDSII. Then object-based classification of these images was carried out. The comparison of the classification results (both qualitative and quantitative) show that the object-based approach gives about 10-15% higher accuracy, much better results in terms of area estimation and change detection of snow covered areas as compared to traditional pixel-based classification. The results also indicate that object based classification is more useful in mountainous regions to avoid confusion among classes produced by shadows."

Kegiatan pemetaan habitat bentik perairan dangkal dimaksudkan untuk mengetahui gambaran mengenai sebaran habitat bentik yang dapat dipakai sebagai acuan dalam perencanaan dan pengembangan suatu kawasan secara optimal dimasa mendatang. Penelitian ini bertujuan memetakan habitat bentik hasil klasifikasi berbasis objek menggunakan citra WorldView-2 dengan algoritma SVM (Supported Vector Machine) dan klasifikasi berdasarkan analisis klaster Bray and Curtis dari data lapangan. Metode penelitian  ini menggunakan klasifikasi dua level yaitu level 1 (reef level) dan level 2 (benthic level). Hasil klasifikasi pada level 1 menghasilkan 3 kelas yang terdiri dari darat, perairan dangkal, dan perairan dalam. Sedangkan level 2  dengan skema klasifikasi klaster Bray and Curtis menghasilkan 8 kelas dan dilakukan penyederhanaan menjadi 5 kelas yang terdiri dari lamun (Lm), pasir (Ps), alga (Al), rubble (R), dan karang (Kr). Uji akurasi peta habitat bentik dengan 5 kelas menghasilkan akurasi keseluruhan sebesar 81,9% dengan nilai kappa 0,783

Improved sensor characteristics are generally assumed to increase the potential accuracy of image classification and information extraction from remote sensing imagery. However, the increase in data volume caused by these improvements raise challenges associated with the selection, storage, and processing of this data, and with the cost-effective and timely analysis of the remote sensing datasets. Previous research has extensively assessed the relevance and impact of spatial, spectral and temporal resolution of satellite data on classification accuracy, but little attention has been given to the impact of radiometric resolution. This study focuses on the role of radiometric resolution on classification accuracy of remote sensing data through different classification experiments over three different sites. The experiments were carried out using fine and low scale radiometric resolution images classified through a bagging classification tree. The classification experiments addressed different aspects of the classification road map, including among others, binary and multiclass classification schemes, spectrally and spatially enhanced images, as well as pixel and objects as units of the classification. In addition, the impact of image radiometric resolution on computational time and the information content in fine- and low-resolution images was also explored. While in certain cases, higher radiometric resolution has led to up to 8% higher classification accuracies compared to lower resolution radiometric data, other results indicate that higher radiometric resolution does not necessarily imply improved classification accuracy. Also, classification accuracy of spectral indices and texture bands is not related so much to the radiometric resolution of the original remote sensing images but rather to their own radiometric resolution. Overall, the results of this study suggest that data selection and classification need not always adhere to the highest possible radiometric resolution.

The European Space Agency (ESA) defines as Earth observation (EO) Level 2 information product a single-date multi-spectral (MS) image corrected for atmospheric, adjacency and topographic effects, stacked with its data-derived scene classification map (SCM), whose thematic map legend includes quality layers cloud and cloud-shadow. ESA EO Level 2 product generation is an inherently ill-posed computer vision (CV) problem never accomplished to date in operating mode by any EO data provider at the ground segment. Herein, it is considered: (I) necessary not sufficient pre-condition for the yet-unaccomplished dependent problems of semantic content-based image retrieval (SCBIR) and semantics-enabled information/knowledge discovery (SEIKD) in multi-source EO big data cubes. (II) Synonym of EO Analysis Ready Data (ARD) format. (III) Equivalent to a horizontal policy for background developments in Space Economy 4.0. In compliance with the GEO-CEOS Quality Assurance Framework for EO Calibration/Validation guidelines, to contribute toward filling an analytic and pragmatic information gap from multi-sensor EO big data to timely, comprehensive and operational EO value-adding information products and services, this work presents an innovative AutoCloud+ CV software toolbox for cloud and cloud-shadow quality layer detection in ESA EO Level 2 product. In vision, spatial information dominates color information. Inspired by this true-fact, the inherently ill-posed AutoCloud+ CV software was conditioned, designed and implemented to be “universal”, meaning fully automated (no human-machine interaction is required), near real-time, robust to changes in input data and scalable to changes in MS imaging sensor’s spatial and spectral resolution specifications.

Advancing technical means – just consider the current satellite systems RapidEye or
TerraSAR-X– as well as new European legislation such as the Water Framework Directive or INSPIRE
bringing new reporting obligations, lead to continuously increasing geoinformation demand by
administrative authorities. Existing geospatial information systems such as the German Authoritative
Topographic and Cartographic Information System (ATKIS) or the Euopean CORINE Land Cover
(CLC) can not fully meet this demand, bringing the focus to current studies and developments to
update and refine these geodata sets. Against this background a consortium of eleven members started
the DeCOVER project in 2006. The project aim is to develop a concept to establish and update land
cover and land use information via geospatial web services based on up-to-date remote sensing information.
This article presents the fundamental aspects and latest results of the DeCOVER project, such as the
sequential processing chain. It focuses on the developed change detection methodology, which is an
integral part of DeCOVER. Different change indicators are implemented based on a comparison of
the input satellite data of different dates. These indicators in combination with a transitionprobability-
matrix are used to limit the new possible classes and control the subsequent recursive
processing to verify the indicated changes according to the sorted probabilities of change. The results
of the proposed object-based change detection process chain are compared to change-detection results
obtained by completely visual interpretation. Finally all results are assembled to a resultant change
indication map.

Object-based image analysis (OBIA) is rapidly emerging as a valuable method for integrating the data processing techniques and GIS approaches classically employed in archaeology. OBIA is intended to replicate human perception by using a protocol of (semi)automated image segmentation and classification. However, the lack of a theoretical background adapted to the specificities of the archaeological discipline is still preventing researchers from finding a shared language and a common protocol of investigation necessary to allow the comparability of the results. This article discusses a series of crucial theoretical issues linked to the incompleteness and the equi-/multifinality of the archaeological record and introduces the core concept of Diachronic Semantic Models (DhSM) as a means to integrate the long-term evolution of the archaeological landscape in the conceptual, digital and real world frameworks of the object-based approach. We also present an assessment of the limits and potential of this method, built from a set of case studies from published and unpublished research. Finally, we propose a general workflow of an Archaeological Object-Based Image Analysis (ArchaeOBIA) project, designed for stimulating the development of an operational routine for object-based applications in archaeology.
…

The new satellites sensors such as GeoEye , WorldView and Ikonos , provide new data for better delineation, detection, and visualization of Land cover changes. Several techniques have been proposed and developed for change detection automatically. The main aim of this research is to assess the use of very high-resolution satellites images in monitoring land cover changes in Egypt. Five change detection techniques have been tested using Ikonos and WorldView images. The techniques considered are image difference, image ratio, principal component analysis, post-classification comparison, and multi-date direct classification. The accuracy of each technique is evaluated via the overall accuracy and kappa coefficient. The results showed that the principal component analysis technique produced more accurate change detection map compared to other methods. Also, it was found that changes from vegetation to urban in the study area are greater than changes from bare soil to vegetation.

While automated feature recognition is still in its infancy in archaeology, within the geosciences recent developments have allowed its application to much more irregular typology, such as the supervised classification of landslides. This success is largely due to the geographical object-based image analysis software TRIMBLE eCognition. In archaeology, this programme has been applied to some extent, yet new additions to the array of available methods require a re-evaluation of its potential for feature recognition. For instance, the ability to integrate LiDAR data and aerial photography has long been desired within archaeology. Additionally, the ability to transfer rulesets for the detection of common features can facilitate data and knowledge-sharing amongst researchers. The case study will present three different automated detection methods; using the well-known eCognition ruleset generation based on cognitive reasoning; self-learning algorithms; and adaptive template matching. These techniques are applied to round barrow detection in the Avebury region in southern England, specifically distinguishing between the known variations of barrow, bank and ditch. Each method is assessed according to its usability for large regions and its potential for detecting variable features and complex shapes. The algorithms are intended to prioritise cognitive aspects of human vision such as elevation, size, shape and texture, using the LiDAR data and aerial photography. It is also stressed that ruleset exchange for generally known features and processes is highly important for mapping large areas across borders and is intrinsically supported by eCognition.

To contribute toward systematic European Space Agency (ESA) Earth observation (EO) Level 2 product generation at the ground segment, requiring a cloud/cloud-shadow quality layer detection in operating mode, but never accomplished to date by any EO data provider, this work presents the research and technological development (RTD) of an innovative AutoCloud+ approach. Designed and implemented to be " universal " , i.e., robust to changes in multi-spectral (MS) imaging sensor's spatial and spectral resolution specifications, AutoCloud+ aims at cloud/cloud-shadow detection in multi-sensor single-date MS imagery without human-machine interaction. Sensitive to spatial topological and spatial non-topological information components in addition to MS color properties in compliance with an object-based image analysis (OBIA) paradigm, AutoCloud+ adopts a hybrid (combined top-down/physical model-based and bottom-up/statistical model-based) CV workflow. The AutoCloud+ software prototype was tested starting from a 10 m resolution Sentinel-2 MS imagery, in comparison with alternative cloud/cloud-shadow detectors in operating mode, specifically, the single-date ESA Sentinel-2 (atmospheric) Correction Prototype Processor (Sen2Cor) and the multi-date CNES/CESBIO/DLR MACCS-ATCOR Joint Algorithm (MAJA). In this difficult test case affected by several error-prone situations, AutoCloud+ ran automatically with high accuracy, qualitatively affected by few false positives and few false negatives. KEYWORDS: color naming, color constancy, object-based image analysis (OBIA), physical and statistical data models, radiometric calibration, spatial topological and spatial non-topological information components, vision. 2 GEOBIA'2018

The study of pre-contact anthropogenic mounded features- earthen mounds, shell heaps, and shell rings - in the American Southeast is stymied by the spotty distribution of systematic surveys across the region. Many extant, yet unidentified, archaeological mound features continue to evade detection due to the heavily forested canopies that occupy large areas of the region, making pedestrian surveys difficult and preventing aerial observation. Object-based image analysis (OBIA) is a tool for analyzing light and radar (LiDAR) data and offers an inexpensive opportunity to address this challenge. Using publicly available LiDAR data from Beaufort County, South Carolina and an OBIA approach that incorporates morphometric classification and statistical template matching, we systematically identify over 160 previously undetected mound features. This result improves our overall knowledge of settlement patterns by providing systematic knowledge about past landscapes.

The primary aim of image data collection is to extract information from imagery and turn scene content into knowledge. In Geographic Information Science (GIScience), the term " object-based image analysis " (OBIA) was tentatively introduced in 2006. When it was re-formulated in 2008 as GEOBIA, the primary focus was set on Earth data-derived applications and the interdisciplinary novelty was emphasized of geospatio-temporal reasoning to cope with massive Earth observation (EO) imagery, where to integrate achievements accomplished by the remote sensing (RS) and CV communities. This article tries to (re-)emphasize the role of space/geospatial in image analysis, flanked by recent trends, such as big data analytics. The following key aspects of multi-source EO image analysis are promoted: (1) EO image enhancement and interoperability/harmonization at the radiometric and semantic levels of analysis; (2) big raster data storage and exploitation, affected by ongoing lacks in tackling spatiotemporal information in vector format; (3) deep (multi-scale) distributed CV systems (artificial neural networks); (4) hybrid (combined deductive/top-down and inductive/bottom-up) inference, (5) structured CV system of systems design, based on a convergence of spatial and color evidence, (6) consistency of CV with human visual perception; (7) semantic content-based image retrieval. The plea for a manifesto GEOBIA 2.0 is backed up backed up by some illustrative examples from the field of GIScience applications.

The objective of this paper is to compare object-based and per-pixel classifiers in a systematic manner using high resolution urban imagery. The prevailing opinion is that object-based methods perform better than single-pixel classifiers, but there has been no formal investigation of this claim using multiple images and identical training samples in a detailed land-cover classification. Furthermore, there has been no standardized study of how different object-based  segmentation and scale parameters improve high resolution urban classifications. We used two subsets of QuickBird over Phoenix and Scottsdale, Arizona, to test these issues. Our results show that small-scale segmentation (10) produces higher accuracy. A combination of equally balanced shape and spectral homogeneity (0.5) with compactness parameter of 0.5 is the most effective for image segmentation. The highest overall accuracy was achieved using a per-pixel Minimum distance classifier, but it was only marginally more accurate than the object-based classification.

Land-use mapping is critical for global change research. In Central Arizona, U.S.A., the spatial distribution of land use is important for sustainable land management decisions. The objective of this study was to create a land-use map that serves as a model for the city of Maricopa, an expanding urban region in the Sun Corridor of Arizona. We use object-based image analysis to map six land-use types from ASTER imagery, and then compare this with two per-pixel classifications. Our results show that a single segmentation, combined with intermediary classifications and merging, morphing, and growing image-objects, can lead to an accurate land-use map that is capable of utilizing both spatial and spectral information. We also employ a moving-window diversity assessment to help with analysis and improve post-classification modifications.

Terrain as a continuum can be categorized into landform units that exhibit common physical and morphological characteristics of land surface which may serve as a boundary condition for a wide range of application domains. However, heterogeneous views, definitions, and applications on landforms yield incompatible nomenclature that lacks interoperability. Yet, there is still room for developing methods for classification of land surface into landforms that can provide different disciplines with a basis of landscape description that is also commonsense to human insight.This study proposes a method of landform classification that reveals general geomorphometry of the landscape. A set of landform classes that are commonsense to human insight and relevant to various disciplines is adopted to generate landforms at the landscape scale. The proposed classification method is based on local geometry of the surface and the geomorphometric context in a higher level framework. A set of digital terrain models (DTMs) at relevant scale is utilized where local geometry is represented with morphometric DTMs, and the geomorphometric context is incorporated through ‘relative terrain position’ and ‘terrain network.’ ‘Object-based image analysis’ tools that have the ability to segment and classify DTMs into representative terrain objects and connect those objects in a multi-level hierarchy is utilized. Ambiguities in landforms both in attribute and geographical space are represented via fuzzy classification. The proposed method is applied to two different case areas to evaluate the efficiency and stability of the outcomes. Results reveal a reasonable amount of consistency where landform classes can be utilized as general or multi-purpose regarding some ambiguity that is already inherent in landforms.

The designation of marble provenance plays an important role in Cultural History, Archeology and Geosciences in general. In the
multidisciplinary approach to explore marble provenance, petrography plays a key role. This paper presents a novel method for automatic
image analysis of marble micrographs: Object Based Image Analysis (OBIA), via the incorporation of petrographic expert
knowledge, enables the reliable extraction of mineral grains and yields a wealth of quantitative shape and texture measures. A work
flow is introduced for extracting mineral shape characteristics from marble micrographs, comprising data acquisition, pre-processing
and Object Based Image Analysis. Therefore verifiable parameters and analysis supply marble provenance research particularly for
multiple sample analysis in an efficient and timely manner.

Die Bestimmung der Herkunft von Marmoren ist von großer Bedeutung für Kulturgeschichte, Archäologie und allgemein für die
Geowissenschaften. Die Petrographie spielt im multidisziplinären Ansatz der Herkunftsbestimmung eine wichtige Rolle. In diesem
Artikel wird eine neue Methode zur automatischen Bildanalyse von Marmor Dünnschliffen präsentiert: Objektbasierte Bildanalyse
(OBIA) kann petrographisches Expertenwissen in den Analysevorgang einbeziehen und ermöglicht so, neben einer verlässlichen
Extraktion von Mineralkörnern, die Darstellung einer Vielzahl von quantitativen Form- und Texturparametern.
Es wird ein Arbeitsablauf zur Extraktion von Formparametern von Calcitkörnern aus Marmordünnschliffen vorgestellt: Datenerfassung,
Datenvorverarbeitung und Objektbasierte Bildanalyse. Die Analyse liefert nachvollziehbare Ergebnisse und unterstützt dadurch
besonders bei Mehrfachprobenauswertung die zeiteffiziente Herkunftsbestimmung von Marmoren.

A changing climate brings increased risks for the tropical Caribbean island of Anguilla. Rising sea levels and increased hurricane frequency and intensity both contribute to beach erosion. With tourism accounting for 64% of gross domestic product in 2011, beach losses equal tourism losses. Yet tourism activities themselves contribute to environmental degradation of the coast. Information is needed on the rate and location of coastal change.

This study used an earth observation approach to overcome the limitations of field-based beach monitoring in Anguilla. Two high resolution costal land cover maps were produced using an object-based image analysis classification approach on fully pre-processed imagery sets. The 2004 classification used Ikonos (1 m) visible imagery and the 2011 classification used WorldView-2 (2 m) visible and near-infrared imagery. Both classifications were assisted by elevation data. Ikonos classification accuracies (85.9–92.5%) approached those of the highly accurate WorldView-2 classifications (97.2%). An observed 72.5% increase in ironshore was largely at the expense of coastal vegetation. A spectral endmember library of sand, water and vegetation, developed from fully pre-processed Landsat ETM+ imagery, showed perfect separation. This library was used in a linear spectral mixture analysis of two Landsat scenes contemporaneous to the high resolution classifications, with good results (RMSE of 0.024 and 0.026).

A correlation analysis evaluated the ability of the linear spectral unmixing to estimate actual land cover proportions. Samples containing only sand, water and vegetation showed strong and significant correlations (Spearman’s rank correlation coefficients of 0.79–0.91 for 2004-5 and 0.84–0.94 for 2011), comparing favourably to similar studies. Fractional land cover estimates from a standardised spectral endmember library could act as a reasonable proxy for actual land cover proportions. Use of freely available, moderate resolution imagery would provide substantial costs savings and increased temporal resolution, better matching the dynamics of coastal change, and overcoming the limitations of field-based monitoring. A semi-automated workflow was suggested, using historical imagery to determine trends in coastal change. This valuable information could be used to project future coastal change, informing coastal management decisions around vulnerability, development planning and conservation.

This paper illustrates the classification of corn in the LiDAR data using Landsat indices as features. The Landsat data are calibrated in order to have a noise free reflectance image. Preliminary field in-situ spectral measurements were carried out in order to determine the unique reflectance values of corn in Butuan City, Philippines. In-situ spectral response measurement was done using Ocean OpticsTM VIS-NIR Spectrometer with spectral range from 350 nm to 1000 nm. The computed average reflectance values of corn samples would then be used in order to derive specific values for the Enhance Vegetation Index (EVI) and Green Ratio of corns within the study area. When RGB aerial images are unavailable, classification using LiDAR data only can proved to be futile. However, the Green Ratio and EVI from the transformed Landsat image could be the key features to classify any entities in the image. This paper highlights the classification of a certain class of vegetation objects having a height of 0.5 to 2 meters in the LiDAR nDSM. This class was referred to as class “Medium Elevation” in this paper. Subclasses of this class are class “Corn” and class “Shrub”. Segmentation was based on LiDAR nDSM and Intensity images. Samples were then collected for each of the two classes. A Supervised learning algorithm called SVM (Support Vector Machine) was used to classify entities in the image. The SVM model was constructed using the LiDAR derivatives (nDSM, DSM intensity), Green Ratio, and EVI as features. High overall accuracies were obtained for the classification of corn and shrub in both the train (91.08%) and test site (91.52%). With this, moderate resolution image like Landsat can indeed compliment other remotely sensed data like LiDAR. Specific band ratios derived from the spectral signature of corn, and features like the Green Ratio and Enhanced Vegetation Index can prove to be reliable features in discriminating corn from other vegetation species.

The analysis of human crowds has widespread uses from law enforcement to urban engineering and traffic management. All of these require a crowd to first be detected, which is the problem addressed in this paper. Given an image, the algorithm we propose segments it into crowd and non-crowd regions. The main idea is to capture two key properties of crowds: (i) on a narrow scale, its basic element should look like a human (only weakly so, due to low resolution, occlusion, clothing variation etc.), while (ii) on a larger scale, a crowd inherently contains repetitive appearance elements. Our method exploits this by building a pyramid of sliding windows and quantifying how "crowd-like" each level of the pyramid is using an underlying statistical model based on quantized SIFT features. The two aforementioned crowd properties are captured by the resulting feature vector of window responses, describing the degree of crowd-like appearance around an image location as the surrounding spatial extent is increased.

Illumination invariance remains the most researched, yet the most challenging aspect of automatic face recognition. In this paper we propose a novel, general recognition framework for efficient matching of individual face images, sets or sequences. The framework is based on simple image processing filters that compete with unprocessed greyscale input to yield a single matching score between individuals. It is shown how the discrepancy between illumination conditions between novel input and the training data set can be estimated and used to weigh the contribution of two competing representations. We describe an extensive empirical evaluation of the proposed method on 171 individuals and over 1300 video sequences with extreme illumination, pose and head motion variation. On this challenging data set our algorithm consistently demonstrated a dramatic performance improvement over traditional filtering approaches. We demonstrate a reduction of 50-75% in recognition error rates, the best performing method-filter combination correctly recognizing 96% of the individuals.

This report describes the design, implementation, evaluation and original enhancements to the Live-Wire method for 2D and 3D image segmentation. Live-Wire 2D employs a semi-automatic paradigm; the user is asked to select a few boundary points of the object to segment, to steer the process in the right direction, while the result is displayed in real time. In our implementation segmentation is extended to three dimensions by performing this process on a slice-by-slice basis. User's time and involvement is further reduced by allowing him to specify object contours in planes orthogonal to the slices. If these planes are chosen strategically, Live-Wire 3D can perform 2D segmentation in the plane of each slice automatically. This report also proposes two improvements to the original method, path heating and a new graph edge feature function based on variance of path properties along the boundary. We show that these improvements lead up to a 33% reduction in interaction with the user, and improved delineation in presence of strong interfering edges.

Detecting informal settlements might be one of the most challenging tasks within urban remote sensing. This phenomenon occurs mostly in developing countries. In order to carry out the urban planning and development tasks necessary to improve living conditions for the poorest world-wide, an adequate spatial data basis is needed
(see Mason, O. S. & Fraser, C. S., 1998). This can only be obtained through the analysis of remote sensing data,
which represents an additional challenge from a technical point of view. Formal settlements by definition are mapped sufficiently for most purposes. However, this does not hold for informal settlements. Due to their microstructure and instability of shape, the detection of these settlements is substantially more difficult. Hence, more sophisticated data and methods of image analysis are necessary, which ideally act as a spatial data basis for a further informal settlement management. While these methods are usually quite labour-intensive, one should nonetheless bear in mind cost-effectivity of the applied methods and tools. In the present article, it will be shown how eCognition can be used to detect and discriminate informal settlements from other land-use-forms
by describing typical characteristics of colour, texture, shape and context. This software is completely objectoriented
and uses a patented, multi-scale image segmentation approach. The generated segments act as image
objects whose physical and contextual characteristics can be described by means of fuzzy logic. The article will
show methods and strategies using eCognition to detect informal settlements from high resolution space-borne
image data such as IKONOS. A final discussion of the results will be given.

Remote sensing from airborne and spaceborne platforms provides valuable data for mapping, environmental
monitoring, disaster management and civil and military intelligence. However, to explore the full value of these data, the
appropriate information has to be extracted and presented in standard format to import it into geo-information systems and
thus allow efficient decision processes. The object-oriented approach can contribute to powerful automatic and semiautomatic
analysis for most remote sensing applications. Synergetic use to pixel-based or statistical signal processing
methods explores the rich information contents. Here, we explain principal strategies of object-oriented analysis, discuss
how the combination with fuzzy methods allows implementing expert knowledge and describe a representative example
for the proposed workflow from remote sensing imagery to GIS. The strategies are demonstrated using the first objectoriented
image analysis software on the market, eCognition, which provides an appropriate link between remote sensing
imagery and GIS.