Hyperspectral Imaging: Techniques for Spectral Detection and Classification, Volume 1

Chein-I Chang

Springer Science & Business Media, Jul 31, 2003 - Computers - 370 pages

Hyperspectral Imaging: Techniques for Spectral Detection and Classification is an outgrowth of the research conducted over the years in the Remote Sensing Signal and Image Processing Laboratory (RSSIPL) at the University of Maryland, Baltimore County. It explores applications of statistical signal processing to hyperspectral imaging and further develops non-literal (spectral) techniques for subpixel detection and mixed pixel classification. This text is the first of its kind on the topic and can be considered a recipe book offering various techniques for hyperspectral data exploitation. In particular, some known techniques, such as OSP (Orthogonal Subspace Projection) and CEM (Constrained Energy Minimization) that were previously developed in the RSSIPL, are discussed in great detail. This book is self-contained and can serve as a valuable and useful reference for researchers in academia and practitioners in government and industry.

Contents

• INTRODUCTION: 1
• 11 BACKGROUND: 2
• 12 OUTLINE OF THE BOOK: 3
• 122 Subpixel Detection: 4
• 123 Mixed Pixel Classification MPC: 5
• 1231 Unconstrained MFC: 7
• 1233 Automatic Mixed Pixel Classification AMPC: 8
• 125 Notations to Be Used in the Book: 10
• HYPERSPECTRAL MEASURES: 13
• HYPERSPECTRAL MEASURES FOR SPECTRAL CHARACTERIZATION: 15
• 211 Spectral Information Measure SIM: 16
• 212 Hidden Markov Model HMMBased Measure: 17
• 22 SPECTRAL SIMILARITY MEASURES: 20
• 222 Spectral Information Divergence SID: 21
• 223 Hidden Markov ModelBased Information Divergence HMMID: 23
• 231 Relative Spectral Discriminatory ProBability RSDPB: 24
• 233 Relative Spectral Discriminatory Entropy RSDE: 25
• 24 EXPERIMENTS: 26
• 242 HYDICE Data: 31
• 25 CONCLUSIONS: 34
• SUBPIXEL DETECTION: 37
• TARGET ABUNDANCECONSTRAINED SUBPIXEL DETECTION PARTIALLY CONSTRAINED LEASTSQUARES METHODS: 39
• 32 LINEAR SPECTRAL MIXTURE MODEL: 40
• 33 ORTHOGONAL SUBSPACE PROJECTION OSP: 41
• 34 SUMTOONE CONSTRAINED LEASTSQUARES METHOD SCLS: 44
• 35 NONNEGATTVELY CONSTRAINED LEASTSQUARES METHOD NCLS: 45
• 36 HYPERSPECTRAL IMAGE EXPERIMENTS: 48
• 37 CONCLUSIONS: 50
• TARGET SIGNATURECONSTRAINED SUBPIXEL DETECTION LINEARLY CONSTRAINED MINIMUM VARIANCE LCMV: 51
• 42 LCMV TARGET DETECTOR: 53
• 421 Constrained Energy Minimization CEM: 54
• 422 TargetConstrained InterferenceMinimized Filter TCIMF: 55
• 43 RELATIONSHIP AMONG OSP CEM AND TCIMF: 56
• 44 A COMPARATIVE ANALYSIS BETWEEN CEM AND TCIMF: 58
• 442 Hyperspectral Image Experiments: 61
• 45 SENSITIVITY OF CEM AND TCIMF TO LEVEL OF TARGET INFORMATION: 63
• 451 Computer Simulations: 64
• 452 Hyperspectral Image Experiments: 67
• 46 REALTIME PROCESSING: 68
• 47 CONCLUSIONS: 71
• AUTOMATIC SUBPIXEL DETECTION UNSUPERVISED SUBPIXEL DETECTION: 73
• 52 UNSUPERVISED VECTOR QUANTIZATION UVQBASED ALGORITHM: 74
• 53 UNSUPERVISED TARGET GENERATION PROCESS UTGP: 75
• 54 UNSUPERVISED NCLS UNCLS ALGORITHM: 78
• 55 EXPERIMENTS: 80
• 56 CONCLUSIONS: 87
• AUTOMATIC SUBPEXEL DETECTION ANOMALY DETECTION: 89
• 62 RXD: 91
• 63 LPTD AND UTD: 94
• 64 RELATIONSHIP BETWEEN CEM AND RXD: 97
• 65 REALTIME PROCESSING: 99
• 66 CONCLUSIONS: 102
• SENSITIVITY OF SUBPIXEL DETECTION: 105
• 72 SENSITIVITY OF TARGET KNOWLEDGE: 107
• 73 SENSITIVITY OF NOISE: 111
• 732 Hyperspectral Image Experiments: 116
• 7322 HYDICE Data: 125
• 74 SENSITIVITY OF ANOMALY DETECTION: 129
• 75 CONCLUSIONS: 137
• UNCONSTRAINED MIXED PIXEL CLASSIFICATION: 139
• UNCONSTRAINED MIXED PIXEL CLASSIFICATION LEASTSQUARES SUBSPACE PROJECTION: 141
• 82 A POSTERIORI OSP: 144
• 822 Target Subspace Projection TSP Classifier: 146
• 823 Oblique Subspace Projection OBSP Classifier: 147
• 824 Unconstrained Maximum Likelihood Estimation Classifier: 148
• 83 ESTIMATION ERROR EVALUATED BY ROC ANALYSIS: 150
• 831 Signature Subspace Projection SSP Classifier: 151
• 832 Oblique Subspace Projection OBSP Classifier: 153
• 841 Computer Simulations: 154
• 842 Hyperspectral Data: 156
• 85 CONCLUSIONS: 159
• A QUANTITATIVE ANALYSIS OF MIXEDTOPURE PIXEL CONVERSION MPCV: 161
• 92 CONVERSION OF MFC TO PPC: 162
• 921 MixedtoPure Pixel Converter MPCV: 163
• 922 Minimum DistanceBased Classification: 164
• 923 Fishers Linear Discriminant Analysis LDA: 166
• 924 Unsupervised Classification: 169
• 94 COMPARATIVE PERFORMANCE ANALYSIS: 171
• 95 CONCLUSIONS: 177
• CONSTRAINED MIXED PIXEL CLASSIFICATION: 179
• TARGET ABUNDANCECONSTRAINED MIXED PIXEL CLASSIFICATION TACMPC: 181
• 102 FULLY CONSTRAINED LEASTSQUARES APPROACH: 183
• 103 MODIFIED FULLY CONSTRAINED LEASTSQUARES MFCLS APPROACH: 184
• 104 COMPUTER SIMULATIONS AND REAL HYPERSPECTRAL IMAGE EXPERIMENTS: 186
• 1042 AVIKIS Image Experiments: 188
• 1043 HYDICE Image Experiments: 193
• 105 NEAR REALTIME IMPLEMENTATION: 201
• 106 CONCLUSIONS: 205
• TARGET SIGNATURECONSTRAINED MIXED PIXEL CLASSIFICATION TSCMPC LCMV CLASSIFIERS: 207
• 112 LCMV CLASSIFIERS: 208
• 113 BOWLES ET ALS FILTER VECTORS FV ALGORITHM: 209
• 114 COLOR ASSIGNMENT OF LCMV CLASSIFIERS: 211
• 115 EXTENSION OF CEM FILTER TO CLASSIFIERS: 213
• 1154 TargetConstrained InterferenceMinimized TCIM Classifier: 214
• 117 HYPERSPECTRAL IMAGE EXPERIMENTS: 218
• 118 REALTIME IMPLEMENTATION FOR LCMV CLASSIFIERS: 223
• 119 CONCLUSIONS: 227
• TARGET SIGNATURECONSTRAINED MIXED PIXEL CLASSIFICATION TSCMPC LINEARLY CONSTRAINED DISCRIMINANT ANALYSIS ...: 229
• 122 LCDA: 230
• 123 WHITENING PROCESS FOR LCDA: 233
• 124 BOWLES ET ALS FILTER VECTORS FV ALGORITHM: 234
• 125 COMPUTER SIMULATIONS AND HYPERSPECTRAL IMAGE EXPERIMENTS: 236
• 126 CONCLUSIONS: 241
• AUTOMATIC MIXED PIXEL CLASSIFICATION AMPC: 243
• AUTOMATIC MIXED PIXEL CLASSIFICATION AMPC UNSUPERVISED MIXED PIXEL CLASSIFICATION: 245
• 132 UNSUPERVISED MFC: 246
• 134 AUTOMATIC TARGET DETECTION AND CLASSIFICATION: 253
• 135 CONCLUSIONS: 255
• AUTOMATIC MIXED PIXEL CLASSIFICATIO AMPC ANOMALY CLASSIFICATION: 257
• 142 TARGET DISCRIMINATION MEASURES: 258
• 143 ANOMALY CLASSIFICATION: 260
• 145 ANALYSIS ON TARGET CORRELATION USING TARGET DISCRIMINATION MEASURES: 265
• 146 ONLINE IMPLEMENTATION: 270
• 147 CONCLUSIONS: 274
• AUTOMATIC MIXED PIXEL CLASSIFICATION AMPC LINEAR SPECTRAL RANDOM MIXTURE ANALYSIS LSRMA: 277
• 152 INDEPENDENT COMPONENT ANALYSIS ICA: 279
• 153 ICABASED LSRMA: 280
• 1531 Relative EntropyBased Measure for ICA: 281
• 1532 Learning Algorithm to Find Separating Matrix W: 282
• 154 EXPERIMENTS: 284
• 1542 HYDICE Image Experiments: 289
• 155 3D ROC ANALYSIS FOR LSRMA: 295
• 156 CONCLUSIONS: 302
• AUTOMATIC MIXED PIXEL CLASSIFICATION AMPC PROJECTION PURSUIT: 305
• 162 PROJECTION PURSUIT: 307
• 163 EVOLUTIONARY ALGORITHM EA: 308
• 164 THRESHOLDING OF PROJECTION IMAGES USING ZERODETECTION: 310
• 165 EXPERIMENTS: 311
• 1652 HYDICE Data Experiments: 313
• 166 CONCLUSIONS: 318
• ESTIMATION FOR VIRTUAL DIMENSIONALITY OF HYPERSPECYRAL IMAGERY: 319
• 172 NEYMANPEARSON DETECTION THEORYBASED EIGEN THRESHOLDING ANALYSIS HFC METHOD: 321
• 173 ESTIMATION OF NOISE COVARIANCE MATRIX: 324
• 1731 Residual Analysis Roger 1996: 325
• SpatialSpectral Prediction Noise Estimation Roger and Arnold 1996: 326
• 174 NOISE ESTIMATIONBASED EIGENTHRESHOLDING: 327
• 175 OTHER METHODS FOR FINDING VD: 329
• 1752 Malinowskis Empirical Indication Function EIF Method: 330
• 1762 AVIRIS and HYDYCE Image Experiments: 333
• 177 CONCLUSIONS: 336
• CONCLUSIONS AND FURTHER TECHNIQUES: 339
• 182 MATHEMATICAL TAXONOMY OF TECHNIQUES: 341
• 183 EXPERIMENTS: 343
• 184 ROC ANALYSIS FOR SUBPIXEL DETECTION AND MIXED PIXEL CLASSIFICATION: 344
• 185 SENSITIVITY ISSUES: 345
• 187 FURTHER TECHNIQUES: 346
• 1872 Convex Cone Analysis: 347
• 1873 Kalman FilterBased Linear Unmixing: 348
• 1875 Band Selection: 349
• 1876 Linear Mixture AnalysisBased Data Compression: 350
• 1877 Radial Basis Function Neural Network Approach: 351
• GLOSSARY: 353
• REFERENCES: 357
• INDEX: 369

Bibliographic information

• Title: Hyperspectral Imaging: Techniques for Spectral Detection and Classification, Volume 1; Hyperspectral Imaging: Techniques for Spectral Detection and Classification, Chein-I. Chang
• Author: Chein-I Chang
• Edition: illustrated
• Publisher: Springer Science & Business Media, 2003
• ISBN: 0306474832, 9780306474835
• Length: 370 pages
• Subjects: Computers / Artificial Intelligence / Computer Vision & Pattern Recognition; Computers / Image Processing; Computers / Optical Data Processing; Computers / Software Development & Engineering / Computer Graphics; Computers / Software Development & Engineering / General; Medical / Allied Health Services / Imaging Technologies; Medical / Biochemistry; Medical / Diagnostic Imaging / General; Medical / Radiology, Radiotherapy & Nuclear Medicine; Science / Chemistry / Environmental; Science / Chemistry / General; Science / Earth Sciences / General; Science / Earth Sciences / Geography; Science / Earth Sciences / Geology; Science / Earth Sciences / Meteorology & Climatology; Science / Environmental Science; Technology & Engineering / Environmental / General; Technology & Engineering / Imaging Systems; Technology & Engineering / Remote Sensing & Geographic Information Systems