Prasad / Bruce / Chanussot Optical Remote Sensing

Advances in Signal Processing and Exploitation Techniques
1. Auflage 2011
ISBN: 978-3-642-14212-3
Verlag: Springer
Format: PDF
Kopierschutz: 1 - PDF Watermark

E-Book, Englisch, Band 3, 344 Seiten

Reihe: Augmented Vision and Reality

Optical remote sensing relies on exploiting multispectral and hyper spectral imagery possessing high spatial and spectral resolutions respectively. These modalities, although useful for most remote sensing tasks, often present challenges that must be addressed for their effective exploitation. This book presents current state-of-the-art algorithms that address the following key challenges encountered in representation and analysis of such optical remotely sensed data. Challenges in pre-processing images, storing and representing high dimensional data, fusing different sensor modalities, pattern classification and target recognition, visualization of high dimensional imagery.

Autoren/Hrsg.

Prasad, Saurabh

Bruce, Lori M.

Chanussot, Jocelyn

Inhaltsverzeichnis

1;Optical Remote Sensing;2
1.1;Preface;4
1.2;Contents;6
1.3;1 Introduction;8
1.3.1;1…Optical Remote Sensing: The Processing Chain;9
1.3.2;2…Optical Remote Sensing: Key Challenges for Signal Processing and Effective Exploitation;11
1.3.3;References;14
1.4;2 Hyperspectral Data Compression Tradeoff;16
1.4.1;Abstract;16
1.4.2;1…Introduction;16
1.4.3;2…Data Acquisition Process and Compression Properties;17
1.4.3.1;2.1 Data Acquisition Process;17
1.4.3.2;2.2 Lossy, Lossless, Near-Lossless;19
1.4.3.3;2.3 Onboard;20
1.4.3.4;2.4 Image Distribution;20
1.4.3.5;2.5 Data Availability;22
1.4.4;3…Trends in Compression Algorithms;22
1.4.4.1;3.1 Prediction-Based;23
1.4.4.2;3.2 Vector Quantization;24
1.4.4.3;3.3 Transform Methods;25
1.4.4.3.1;3.3.1 Transform;25
1.4.4.3.2;3.3.2 Coding;26
1.4.4.4;3.4 Lossy to Lossless;28
1.4.4.5;3.5 What is in Use Now?;29
1.4.5;4…Ensuring Sufficient Quality;29
1.4.5.1;4.1 Why Bothering with Lossy Compression?;29
1.4.5.2;4.2 Quality Evaluation;30
1.4.5.3;4.3 Making Comparison Easier;32
1.4.6;5…Reference Results;32
1.4.7;6…Conclusion;34
1.4.8;Acknowledgments;34
1.4.9;References;34
1.5;3 Reconstructions from Compressive Random Projections of Hyperspectral Imagery;37
1.5.1;Abstract;37
1.5.2;1…Introduction;38
1.5.3;2…Compressive-Projection Principal Component Analysis (CPPCA);39
1.5.3.1;2.1 Overview of CPPCA;40
1.5.3.2;2.2 The CPPCA Algorithm;41
1.5.3.2.1;2.2.1 Eigenvector Recovery;42
1.5.3.2.2;2.2.2 Coefficient Recovery;43
1.5.4;3…Compressed Sensing (CS);44
1.5.5;4…Empirical Comparisons on Hyperspectral Imagery;45
1.5.5.1;4.1 Performance of Single-Task and Multi-Task CS;46
1.5.5.2;4.2 Performance of CPPCA and CS;48
1.5.5.3;4.3 Execution Times;51
1.5.6;5…Conclusions;52
1.5.7;References;52
1.6;4 Integrated Sensing and Processing for Hyperspectral Imagery;55
1.6.1;Abstract;55
1.6.2;1…Introduction;56
1.6.3;2…Variable Resolution Hyperspectral Sensing;58
1.6.3.1;2.1 Mathematical Representation;58
1.6.3.2;2.2 Reduced Resolution Imaging;61
1.6.4;3…Experimental Results;65
1.6.4.1;3.1 Improving SNR Using Hadamard Multiplexing;65
1.6.4.2;3.2 Variable Resolution Hyperspectral Sensing;67
1.6.5;4…Summary;69
1.6.6;References;70
1.7;5 Color Science and Engineering for the Display of Remote Sensing Images;71
1.7.1;Abstract;71
1.7.2;1…Introduction;71
1.7.3;2…Challenges;72
1.7.3.1;2.1 Information Loss;72
1.7.3.2;2.2 Metrics;73
1.7.3.3;2.3 Visual Interpretation;73
1.7.3.4;2.4 Color Saturation and Neutrals;75
1.7.3.5;2.5 Color Blindness;76
1.7.3.6;2.6 Case Study: Principal Components Analysis for Visualization;76
1.7.4;3…Some Solutions;77
1.7.4.1;3.1 Optimized Basis Functions;78
1.7.4.2;3.2 Adapting Basis Functions;80
1.7.4.3;3.3 White Balance;83
1.7.5;4…Conclusions and Open Questions;84
1.7.6;References;85
1.8;6 An Evaluation of Visualization Techniques for Remotely Sensed Hyperspectral Imagery;86
1.8.1;Abstract;86
1.8.2;1…Introduction;86
1.8.3;2…Image Construction;88
1.8.4;3…Comparative Visualization Techniques;89
1.8.4.1;3.1 Hard Classification Visualization;90
1.8.4.2;3.2 Soft Classification Visualization;90
1.8.4.3;3.3 Double Layer Visualization;91
1.8.5;4…Experimental Design and Settings;92
1.8.6;5…Experimental Tasks and Results;94
1.8.6.1;5.1 Global Pattern Display Capability;94
1.8.6.1.1;5.1.1 Perceptual Edge Detection;94
1.8.6.1.1.1;Task;95
1.8.6.1.1.2;Results;95
1.8.6.1.2;5.1.2 Block Value Estimation;96
1.8.6.1.2.1;Task;96
1.8.6.1.2.2;Result;97
1.8.6.2;5.2 Ability to Convey Local Information;97
1.8.6.2.1;5.2.1 Class Recognition;98
1.8.6.2.1.1;Task;98
1.8.6.2.1.2;Results;98
1.8.6.2.2;5.2.2 Target Value Estimation;100
1.8.6.2.2.1;Task;100
1.8.6.2.2.2;Results;100
1.8.7;6…Discussion and Conclusions;101
1.8.8;Acknowledgments;102
1.8.9;References;102
1.9;7 A Divide-and-Conquer Paradigm for Hyperspectral Classification and Target Recognition;104
1.9.1;Abstract;104
1.9.2;1…Introduction;105
1.9.3;2…The Proposed Framework;107
1.9.3.1;2.1 Subspace Identification: Partitioning the Hyperspectral Space;108
1.9.3.2;2.2 Pre-processing at the Subspace Level;111
1.9.3.2.1;2.2.1 Linear Discriminant Analysis (LDA);111
1.9.3.2.2;2.2.2 Kernel Discriminant Analysis (KDA);112
1.9.3.3;2.3 Classifier;114
1.9.3.4;2.4 Decision Fusion;115
1.9.4;3…Experimental Hyperspectral Datasets;116
1.9.4.1;3.1 Handheld Hyperspectral Data;116
1.9.4.2;3.2 Airborne Hyperspectral Data;117
1.9.5;4…Experimental Setup and Results;118
1.9.5.1;4.1 Experiments with Handheld HSI Data;119
1.9.5.1.1;4.1.1 Experiment 1: MCDF with LDA Based Pre-processing at the Subspace Level;119
1.9.5.1.2;4.1.2 Experiment 2: MCDF with KDA Based Pre-processing at the Subspace Level;120
1.9.5.2;4.2 Experiments with Aerial HSI Data;123
1.9.6;5…Conclusions, Caveats and Future Work;124
1.9.7;Acknowledgments;125
1.9.8;References;125
1.10;8 The Evolution of the Morphological Profile: from Panchromatic to Hyperspectral Images;128
1.10.1;Abstract;128
1.10.2;1…Introduction;129
1.10.3;2…Preliminaries of Mathematical Morphology;130
1.10.3.1;2.1 Fundamental Properties;130
1.10.3.2;2.2 Opening and Closing by Reconstruction;131
1.10.3.3;2.3 Attribute Filters;133
1.10.4;3…Morphological Profiles for the Analysis of Panchromatic Images;136
1.10.4.1;3.1 Morphological Profiles;136
1.10.4.2;3.2 Attribute Profiles;138
1.10.4.3;3.3 Experimental Results and Discussion;139
1.10.5;4…Extended Morphological Profiles to the Analysis of Multispectral and Hyperspectral Images;142
1.10.5.1;4.1 Problem of Extending the Morphological Operators to Multi-tone Images;142
1.10.5.2;4.2 Extended Morphological Profile;143
1.10.5.3;4.3 Extended Attribute Profiles;144
1.10.5.4;4.4 Experimental Results and Discussion;145
1.10.6;5…Conclusion;148
1.10.7;Acknowledgments;149
1.10.8;References;149
1.11;9 Decision Fusion of Multiple Classifiers for Vegetation Mapping and Monitoring Applications by Means of Hyperspectral Data;152
1.11.1;Abstract;152
1.11.2;1…Introduction;153
1.11.3;2…Study Area;154
1.11.3.1;2.1 Image Data;154
1.11.3.2;2.2 Field Data;155
1.11.3.3;2.3 Training Data;155
1.11.4;3…Methods for Fusion of Multiple Classifiers;157
1.11.4.1;3.1 Decision Fusion Using Hierarchical Tree Structure;157
1.11.4.2;3.2 Decision Fusion Using the Hierarchical Tree and Class Membership Values;159
1.11.4.3;3.3 Class-Dependent Neural Networks Ensemble;160
1.11.5;4…Accuracy Assessment;161
1.11.5.1;4.1 Comparison of Classification Results;162
1.11.6;5…Results;163
1.11.6.1;5.1 Results of Various Tested Classifiers;163
1.11.6.2;5.2 Results of Class Dependent Neural Networks;168
1.11.6.3;5.3 Results of Decision Fusion Using Hierarchical Tree Structure;169
1.11.6.4;5.4 Results of Hierarchical Tree Coupled with Probability Labels;171
1.11.6.5;5.5 The Assessment of Significance of the Accuracy Values;172
1.11.7;6…Conclusions;173
1.11.8;Acknowledgments;174
1.11.9;References;174
1.12;10 A Review of Kernel Methods in Remote Sensing Data Analysis;176
1.12.1;Abstract;176
1.12.2;1…Introduction;177
1.12.2.1;1.1 Classification with Kernels;177
1.12.2.2;1.2 Model Inversion with Kernels;178
1.12.2.3;1.3 Feature Extraction with Kernels;178
1.12.3;2…Introduction to Kernel Methods;179
1.12.3.1;2.1 Measuring Similarity with Kernels;179
1.12.3.2;2.2 Positive Definite Kernels;179
1.12.3.3;2.3 Basic Operations with Kernels;180
1.12.3.4;2.4 Standard Kernels;180
1.12.3.5;2.5 Kernel Development;181
1.12.4;3…Kernel Methods in Remote Sensing Data Classification;182
1.12.4.1;3.1 Support Vector Machine;182
1.12.4.2;3.2 nu -Support Vector Machine;183
1.12.4.3;3.3 Support Vector Data Description;185
1.12.4.4;3.4 One-Class Support Vector Machine;185
1.12.4.5;3.5 Kernel Fisher’s Discriminant;186
1.12.4.6;3.6 Experimental Results for Supervised Classification;187
1.12.4.6.1;3.6.1 Linear versus nonlinear;188
1.12.4.6.2;3.6.2 {\varvec \nu} -SVM versus OC-SVM;188
1.12.4.6.3;3.6.3 Support Vector versus Fisher’s Discriminant;189
1.12.4.7;3.7 Semisupervised Image Classification;189
1.12.4.7.1;3.7.1 Manifold-Based Regularization Framework;190
1.12.4.7.2;3.7.2 Semisupervised Regularization Framework;190
1.12.4.7.3;3.7.3 Laplacian Support Vector Machine;191
1.12.4.7.4;3.7.4 Transductive SVM;192
1.12.4.8;3.8 Experimental Results for Semisupervised Classification;192
1.12.5;4…Kernel Methods in Biophysical Parameter Estimation;193
1.12.5.1;4.1 Support Vector Regression;194
1.12.5.2;4.2 Relevance Vector Machines;195
1.12.5.3;4.3 Gaussian Processes;197
1.12.5.4;4.4 Experimental Results;198
1.12.6;5…Kernel Methods for Feature Extraction;199
1.12.6.1;5.1 Mutivariate Analysis Methods;200
1.12.6.1.1;5.1.1 Principal Component Analysis;200
1.12.6.1.2;5.1.2 Partial Least Squares;201
1.12.6.2;5.2 Kernel Multivariate Analysis;201
1.12.6.2.1;5.2.1 Kernel Principal Component Analysis;202
1.12.6.2.2;5.2.2 Kernel Partial Least Squares;203
1.12.6.3;5.3 Experimental Results;203
1.12.7;6…Future Trends in Remote Sensing Kernel Learning;204
1.12.7.1;6.1 Multiple Kernel Learning;204
1.12.7.2;6.2 Transfer Learning;204
1.12.7.3;6.3 Structured Learning;205
1.12.7.4;6.4 Active Learning;205
1.12.7.5;6.5 Parallel Implementations;205
1.12.8;7…Conclusions;206
1.12.9;Acknowledgments;206
1.12.10;References;206
1.13;11 Exploring Nonlinear Manifold Learning for Classification of Hyperspectral Data;212
1.13.1;Abstract;212
1.13.2;1…Introduction;213
1.13.3;2…Nonlinear Manifold Learning for Dimensionality Reduction;214
1.13.3.1;2.1 Dimensionality Reduction Within a Graph Embedding Framework;215
1.13.3.2;2.2 Global Manifold Learning;215
1.13.3.2.1;2.2.1 Isometric Feature Mapping (Isomap);216
1.13.3.2.2;2.2.2 Kernel Principal Component Analysis (KPCA);216
1.13.3.3;2.3 Local Manifold Learning;217
1.13.3.3.1;2.3.1 Locally Linear Embedding (LLE);217
1.13.3.3.2;2.3.2 Local Tangent Space Alignment (LTSA);218
1.13.3.3.3;2.3.3 Laplacian Eigenmaps (LE);218
1.13.3.4;2.4 Supervised Local Manifold Learning;219
1.13.3.5;2.5 Kernel-Based Out-of-Sample Extension;219
1.13.4;3…Remotely Sensed Data for Comparative Experiments;220
1.13.4.1;3.1 Botswana Hyperion Data (BOT);220
1.13.4.2;3.2 Kennedy Space Center AVIRIS Data (KSC);221
1.13.4.3;3.3 Indian Pine AVIRIS Data (IND PINE);222
1.13.4.4;3.4 ACRE ProspectTIR Data (ACRE);222
1.13.5;4…Experimental Results;222
1.13.5.1;4.1 Performance of Dimensionality Reduction Methods (DR) for BOT Hyperion Data;223
1.13.5.2;4.2 Comparison of DR Methods for BOT, KSC, IND PINE, and ACRE Sites;227
1.13.5.3;4.3 Manifold Coordinates for DR Methods;231
1.13.6;5…Summary and Conclusions;234
1.13.7;Acknowledgment;237
1.13.8;References;237
1.14;12 Recent Developments in Endmember Extraction and Spectral Unmixing;240
1.14.1;Abstract;240
1.14.2;1…Introduction;241
1.14.3;2…Linear Spectral Unmixing;243
1.14.3.1;2.1 Problem Formulation;243
1.14.3.2;2.2 Endmember Extraction;244
1.14.3.2.1;2.2.1 N-FINDR;245
1.14.3.2.2;2.2.2 Orthogonal Subspace Projection (OSP);246
1.14.3.2.3;2.2.3 Vertex Component Analysis (VCA);247
1.14.3.2.4;2.2.4 Automatic Morphological Endmember Extraction (AMEE);247
1.14.3.2.5;2.2.5 Spatial Spectral Endmember Extraction (SSEE);248
1.14.3.2.6;2.2.6 Spatial Pre-Processing (SPP);250
1.14.3.3;2.3 Unconstrained Versus Constrained Linear Spectral Unmixing;251
1.14.4;3…Nonlinear Spectral Unmixing;252
1.14.4.1;3.1 Problem Formulation;252
1.14.4.2;3.2 Neural Network-Based Spectral Unmixing;253
1.14.4.3;3.3 Automatic Selection and Labeling of Training Samples;255
1.14.5;4…Experimental Results;256
1.14.5.1;4.1 First Experiment: AVIRIS Hyperspectral Data;256
1.14.5.2;4.2 Second Experiment: DAIS 7915 and ROSIS Hyperspectral Data;259
1.14.5.2.1;4.2.1 Data Description;260
1.14.5.2.2;4.2.2 Fractional Abundance Estimation Results;262
1.14.6;5…Parallel Implementation Case Study;264
1.14.7;6…Conclusions and Future Research;268
1.14.8;Acknowledgements;269
1.14.9;References;269
1.15;13 Change Detection in VHR Multispectral Images: Estimation and Reduction of Registration Noise Effects;273
1.15.1;Abstract;273
1.15.2;1…Introduction;274
1.15.3;2…Notation and Background;275
1.15.4;3…Analysis of Registration Noise Properties;277
1.15.4.1;3.1 Experimental Setup;278
1.15.4.1.1;3.1.1 Experiment 1: Effects of Increasing Misregistration on Unchanged Pixels;279
1.15.4.1.2;3.1.2 Experiment 2: Effects of Increasing Misregistration on Changed Pixels;280
1.15.4.1.3;3.1.3 Experiment 3: Effects of Misregistration at Different Scales;280
1.15.4.2;3.2 Properties of RN in VHR Images;282
1.15.5;4…Proposed Technique for the Adaptive Estimation of the Registration Noise Distribution;287
1.15.6;5…Proposed Change-Detection Technique Robust to Registration Noise;290
1.15.6.1;5.1 Registration Noise Identification;291
1.15.6.2;5.2 Context-Sensitive Decision Strategy for the Generation of the Final Change-Detection Map;293
1.15.7;6…Experimental Results;294
1.15.7.1;6.1 Data Set Description;294
1.15.7.2;6.2 Estimation Results;296
1.15.7.3;6.3 Change-Detection Results;298
1.15.8;7…Discussion and Conclusion;301
1.15.9;References;302
1.16;14 Effects of the Spatial Enhancement of Hyperspectral Images on the Distribution of Spectral Classes;304
1.16.1;Abstract;304
1.16.2;1…Introduction;304
1.16.3;2…Spatial Enhancement of Hyperspectral Images;306
1.16.3.1;2.1 Component Substitution Methods;306
1.16.3.2;2.2 Multiresolution Methods;307
1.16.3.3;2.3 Selected Methods for Testing on HS+Pan Images;308
1.16.3.4;2.4 Evaluation of Spatial Enhancement Methods;310
1.16.4;3…Dimensionality Reduction for the Assessment of Pan-Sharpening Algorithms;310
1.16.5;4…Experimental Results;313
1.16.5.1;4.1 ‘Visual’ Approach;313
1.16.5.1.1;4.1.1 GIHS;313
1.16.5.1.2;4.1.2 HPF;314
1.16.5.1.3;4.1.3 HPF-P;314
1.16.5.1.4;4.1.4 GMMSE;316
1.16.5.2;4.2 Linear and Non-Linear Sample Similarity Measures;316
1.16.5.3;4.3 PCA;318
1.16.5.4;4.4 Kernel PCA;321
1.16.5.5;4.5 Linearity Preserving Projection (LPP);324
1.16.6;5…Conclusions;328
1.16.7;References;328
1.17;15 Fusion of Optical and SAR Data for Seismic Vulnerability Mapping of Buildings;331
1.17.1;Abstract;331
1.17.2;1…Introduction;332
1.17.3;2…Fusion of Optical and Radar Data;332
1.17.4;3…Data Fusion for Vulnerability Assessment;334
1.17.4.1;3.1 The Aim;334
1.17.4.2;3.2 Remote Sensing as a Tool;335
1.17.4.3;3.3 Decision-Level Fusion;340
1.17.5;4…Conclusions;341
1.17.6;Acknowledgments;342
1.17.7;References;342