Since the publishing of the first edition of Classification Methods for Remotely Sensed Data in 2001, the field of pattern recognition has expanded in many new directions that make use of new technologies to capture data and more powerful computers to mine and process it. What seemed visionary but a decade ago is now being put to use and refined in commercial applications as well as military ones. Keeping abreast of these new developments, Classification Methods for Remotely Sensed Data, Second Edition provides a comprehensive and up-to-date review of the entire field of classification methods applied to remotely sensed data. This second edition provides seven fully revised chapters and two new chapters covering support vector machines (SVM) and decision trees. It includes updated discussions and descriptions of Earth observation missions along with updated bibliographic references. After an introduction to the basics, the text provides a detailed discussion of different approaches to image classification, including maximum likelihood, fuzzy sets, and artificial neural networks. This cutting-edge resource: Presents a number of approaches to solving the problem of allocation of data to one of several classes Covers potential approaches to the use of decision trees Describes developments such as boosting and random forest generation Reviews lopping branches that do not contribute to the effectiveness of the decision trees Complete with detailed comparisons, experimental results, and discussions for each classification method introduced, this book will bolster the work of researchers and developers by giving them access to new developments. It also provides students with a solid foundation in remote sensing data classification methods.
Author(s): Brandt Tso; Paul M. Mather
Edition: 2
Publisher: CRC Press
Year: 2009
Language: English
Pages: 367
Tags: Приборостроение;Обработка сигналов;
CLASSIFICATION METHODS FOR REMOTELY SENSED DATA ......Page 1
Title page ......Page 4
Copyright page ......Page 5
Contents ......Page 6
Preface to the Second Edition ......Page 12
Preface to the First Edition ......Page 14
Author Biographies ......Page 20
1. Remote Sensing in the Optical and Microwave Regions ......Page 21
1.1 Introduction to Remote Sensing ......Page 24
1.1.2 Surface Material Reflectance ......Page 25
1.1.3 Spatial and Radiometric Resolution ......Page 28
1.2 Optical Remote Sensing Systems ......Page 30
1.3 Atmospheric Correction ......Page 31
1.3.1 Dark Object Subtraction ......Page 32
1.3.2.1 Modeling the Atmospheric Effect ......Page 33
1.3.2.2 Steps in Atmospheric Correction ......Page 37
1.4 Correction for Topographic Effects ......Page 39
1.5 Remote Sensing in the Microwave Region ......Page 42
1.6 Radar Fundamentals ......Page 43
1.6.1 SLAR Image Resolution ......Page 44
1.6.2 Geometric Effects on Radar Images ......Page 46
1.6.3.1 Surface Roughness ......Page 49
1.6.3.3 Parameters of the Radar Equation ......Page 50
1.7 Imaging Radar Polarimetry ......Page 51
1.7.1 Radar Polarization State ......Page 52
1.7.2 Polarization Synthesis ......Page 54
1.7.3 Polarization Signatures ......Page 55
1.8.1 Multilook Processing ......Page 57
1.8.2 Filters for Speckle Suppression ......Page 58
2. Pattern Recognition Principles ......Page 61
2.1 Feature Space Manipulation ......Page 62
2.1.1 Tasseled Cap Transform ......Page 65
2.1.2 Principal Components Analysis ......Page 66
2.1.3 Minimum/Maximum Autocorrelation Factors (MAF) ......Page 70
2.1.4 Maximum Noise Fraction Transformation ......Page 71
2.2 Feature Selection ......Page 72
2.3.1.1 The k-means Algorithm ......Page 74
2.3.1.2 Fuzzy Clustering ......Page 76
2.3.2.2 Minimum Distance Classifier ......Page 77
2.3.2.3 Maximum Likelihood Classifier ......Page 78
2.4 Combining Classifiers ......Page 81
2.5 Incorporation of Ancillary Information ......Page 82
2.5.2 Using Ancillary Multisource Data ......Page 83
2.6 Sampling Scheme and Sample Size ......Page 85
2.6.1 Sampling Scheme ......Page 86
2.6.2 Sample Size, Scale, and Spatial Variability ......Page 87
2.7 Estimation of Classification Accuracy ......Page 89
2.8 Epilogue ......Page 94
3.1 Multilayer Perceptron ......Page 97
3.1.1 Back-Propagation ......Page 98
3.1.2 Parameter Choice, Network Architecture, and Input/Output Coding ......Page 102
3.1.3 Decision Boundaries in Feature Space ......Page 104
3.1.4 Overtraining and Network Pruning ......Page 108
3.2.1 SOM Network Construction and Training ......Page 110
3.2.1.1 Unsupervised Training ......Page 111
3.2.1.2 Supervised Training ......Page 113
3.2.2 Examples of Self-Organization ......Page 114
3.3 Counter-Propagation Networks ......Page 118
3.3.1 Counter-Propagation Network Training ......Page 119
3.4 Hopfield Networks ......Page 121
3.4.2 Hopfield Network Dynamics ......Page 122
3.4.3 Network Convergence ......Page 123
3.4.4 Issues Relating to Hopfield Networks ......Page 125
3.4.5 Energy and Weight Coding: An Example ......Page 126
3.5 Adaptive Resonance Theory (ART) ......Page 128
3.5.1 Fundamentals of the ART Model ......Page 129
3.5.2 Choice of Parameters ......Page 132
3.5.3 Fuzzy ARTMAP ......Page 133
3.6.1 An Overview ......Page 136
3.6.2 A Comparative Study ......Page 139
4. Support Vector Machines ......Page 145
4.1.1 The Separable Case ......Page 146
4.1.2 The Nonseparable Case ......Page 149
4.2.1 Nonlinear SVMs ......Page 150
4.2.2 Kernel Functions ......Page 152
4.3 Parameter Determination ......Page 155
4.3.1 t-Fold Cross-Validations ......Page 157
4.3.2 Bound on Leave-One-Out Error ......Page 158
4.3.3 Grid Search ......Page 160
4.3.4 Gradient Descent Method ......Page 162
4.4.1 One-against-One, One-against-Others, and DAG ......Page 164
4.4.2.1 Vapnik’s Approach ......Page 166
4.4.2.2 Methodology of Crammer and Singer ......Page 167
4.5 Feature Selection ......Page 169
4.6 SVM Classification of Remotely Sensed Data ......Page 170
4.7 Concluding Remarks ......Page 173
5.1 Introduction to Fuzzy Set Theory ......Page 175
5.1.1 Fuzzy Sets: Definition ......Page 176
5.1.2 Fuzzy Set Operations ......Page 177
5.2 Fuzzy C-Means Clustering Algorithm ......Page 179
5.3 Fuzzy Maximum Likelihood Classification ......Page 182
5.4 Fuzzy Rule Base ......Page 184
5.4.1 Fuzzification ......Page 185
5.4.2 Inference ......Page 189
5.4.3 Defuzzification ......Page 191
5.5.1 Introductory Methodology ......Page 193
5.5.2 Experimental Results ......Page 198
6. Decision Trees ......Page 203
6.1 Feature Selection Measures for Tree Induction ......Page 204
6.1.1 Information Gain ......Page 205
6.1.2 Gini Impurity Index ......Page 208
6.2.1 ID3 ......Page 209
6.2.2 C4.5 ......Page 213
6.2.3 SEE5.0 ......Page 216
6.3 CHAID ......Page 217
6.4 CART ......Page 218
6.5.1 Split Point Selection ......Page 221
6.5.2 Attribute Selection ......Page 223
6.6 Tree Induction from Artificial Neural Networks ......Page 224
6.7 Pruning Decision Trees ......Page 225
6.7.2 Pessimistic Error Pruning (PEP) ......Page 227
6.7.3 Error-Based Pruning (EBP) ......Page 228
6.7.4 Cost Complexity Pruning (CCP) ......Page 229
6.7.5 Minimal Error Pruning (MEP) ......Page 232
6.8.1 Boosting ......Page 234
6.8.2 Random Forest ......Page 235
6.9 Decision Trees in Remotely Sensed Data Classification ......Page 237
6.10 Concluding Remarks ......Page 240
7. Texture Quantization ......Page 241
7.1 Fractal Dimensions ......Page 242
7.1.1 Introduction to Fractals ......Page 243
7.1.2 Estimation of the Fractal Dimension ......Page 244
7.1.2.1 Fractal Brownian Motion (FBM) ......Page 245
7.1.2.2 Box-Counting Methods and Multifractal Dimension ......Page 246
7.2.1 Fourier Power Spectrum ......Page 251
7.2.2 Wavelet Transform ......Page 255
7.3.1 Introduction to the GLCM ......Page 259
7.3.2 Texture Features Derived from the GLCM ......Page 261
7.4.1 MAR Model: Definition ......Page 263
7.4.2 Estimation of the Parameters of the MAR Model ......Page 265
7.5 The Semivariogram and Window Size Determination ......Page 266
7.6.1 Test Image Generation ......Page 269
7.6.2.4 Gray-Level Co-Occurrence Matrix ......Page 270
7.6.3 Segmentation Results ......Page 271
7.6.4 Texture Measure of Remote Sensing Patterns ......Page 272
8. Modeling Context Using Markov Random Fields ......Page 275
8.1.1 Markov Random Fields ......Page 276
8.1.2 Gibbs Random Fields ......Page 277
8.1.3 MRF-GRF Equivalence ......Page 279
8.1.4 Simplified Form of MRF ......Page 281
8.1.5 Generation of Texture Patterns Using MRF ......Page 283
8.2 Posterior Energy for Image Classification ......Page 284
8.3 Parameter Estimation ......Page 287
8.3.1 Least Squares Fit Method ......Page 288
8.3.2 Results of Parameter Estimations ......Page 291
8.4 MAP-MRF Classification Algorithms ......Page 293
8.4.1 Iterated Conditional Modes ......Page 294
8.4.2 Simulated Annealing ......Page 295
8.4.3 Maximizer of Posterior Marginals ......Page 297
8.5 Experimental Results ......Page 298
9. Multisource Classification ......Page 303
9.1.1 Image Fusion Methods ......Page 304
9.1.2 Assessment of Fused Image Quality in the Spectral Domain ......Page 307
9.2 Multisource Classification Using the Stacked-Vector Method ......Page 308
9.3.1 An Overview ......Page 310
9.3.1.1 Feature Extraction ......Page 311
9.3.2 Bayesian Multisource Classification Mechanism ......Page 312
9.3.3 A Refined Multisource Bayesian Model ......Page 314
9.3.4 Multisource Classification Using the Markov Random Field ......Page 315
9.3.5 Assumption of Intersource Independence ......Page 316
9.4.1 Concept Development ......Page 317
9.4.2 Belief Function and Belief Interval ......Page 319
9.4.3 Evidence Combination ......Page 322
9.5 Dealing with Source Reliability ......Page 324
9.5.2 Use of Class Separability ......Page 325
9.5.3 Data Information Class Correspondence Matrix ......Page 326
9.5.4 The Genetic Algorithm ......Page 327
9.6 Experimental Results ......Page 329
Bibliography ......Page 337
