Radar Remote Sensing of Urban Areas (eBook)

Preface 6
Contents 8
1 Review of Radar Remote Sensing on Urban Areas 17
1.1 Introduction 17
1.2 Basics 18
1.2.1 Imaging Radar 19
1.2.2 Mapping of 3d Objects 24
1.3 2d Approaches 27
1.3.1 Pre-processing and Segmentation of Primitive Objects 27
1.3.2 Classification of Single Images 29
1.3.2.1 Detection of Settlements 30
1.3.2.2 Characterization of Settlements 31
1.3.3 Classification of Time-Series of Images 32
1.3.4 Road Extraction 33
1.3.4.1 Recognition of Roads and of Road Networks 33
1.3.4.2 Benefit of Multi-aspect SAR Images for Road Network Extraction 35
1.3.5 Detection of Individual Buildings 36
1.3.6 SAR Polarimetry 36
1.3.6.1 Basics 37
1.3.6.2 SAR Polarimetry for Urban Analysis 39
1.3.7 Fusion of SAR Images with Complementing Data 40
1.3.7.1 Image Registration 40
1.3.7.2 Fusion for Land Cover Classification 41
1.3.7.3 Feature-Based Fusion of High-Resolution Data 42
1.4 3d Approaches 42
1.4.1 Radargrammetry 43
1.4.1.1 Single Image 43
1.4.1.2 Stereo 44
1.4.1.3 Image Fusion 45
1.4.2 SAR Interferometry 45
1.4.2.1 InSAR Principle 45
1.4.2.2 Analysis of a Single SAR Interferogram 48
1.4.2.3 Multi-image SAR Interferometry 50
1.4.2.4 Multi-aspect InSAR 50
1.4.3 Fusion of InSAR Data and Other Remote Sensing Imagery 52
1.4.4 SAR Polarimetry and Interferometry 53
1.5 Surface Motion 54
1.5.1 Differential SAR Interferometry 54
1.5.2 Persistent Scatterer Interferometry 55
1.6 Moving Object Detection 56
References 57
2 Rapid Mapping Using Airborne and Satellite SAR Images 64
2.1 Introduction 64
2.2 An Example Procedure 66
2.2.1 Pre-processing of the SAR Images 66
2.2.2 Extraction of Water Bodies 67
2.2.3 Extraction of Human Settlements 68
2.2.4 Extraction of the Road Network 69
2.2.5 Extraction of Vegetated Areas 71
2.2.6 Other Scene Elements 72
2.3 Examples on Real Data 72
2.3.1 The Chengdu Case 73
2.3.2 The Luojiang Case 76
2.4 Conclusions 79
References 81
3 Feature Fusion Based on Bayesian Network Theory for Automatic Road Extraction 84
3.1 Introduction 84
3.2 Bayesian Network Theory 85
3.3 Structure of a Bayesian Network 87
3.3.1 Estimating Continuous Conditional Probability Density Functions 91
3.3.2 Discrete Conditional Probabilities 94
3.3.3 Estimating the A-Priori Term 95
3.4 Experiments 96
3.5 Discussion and Conclusion 97
References 100
4 Traffic Data Collection with TerraSAR-Xand Performance Evaluation 102
4.1 Motivation 102
4.2 SAR Imaging of Stationary and Moving Objects 103
4.3 Detection of Moving Vehicles 108
4.3.1 Detection Scheme 109
4.3.2 Integration of Multi-temporal Data 111
4.4 Matching Moving Vehicles in SAR and Optical Data 113
4.4.1 Matching Static Scenes 113
4.4.2 Temporal Matching 115
4.5 Assessment 116
4.5.1 Accuracy of Reference Data 116
4.5.2 Accuracy of Vehicle Measurements in SAR Images 118
4.5.3 Results of Traffic Data Collectionwith TerraSAR-X 118
4.6 Summary and Conclusion 122
References 122
5 Object Recognition from Polarimetric SAR Images 124
5.1 Introduction 124
5.2 SAR Polarimetry 126
5.3 Features and Operators 132
5.4 Object Recognition in PolSAR Data 139
5.5 Concluding Remarks 144
References 145
6 Fusion of Optical and SAR Images 147
6.1 Introduction 147
6.2 Comparison of Optical and SAR Sensors 149
6.2.1 Statistics 150
6.2.2 Geometrical Distortions 151
6.3 SAR and Optical Data Registration 152
6.3.1 Knowledge of the Sensor Parameters 152
6.3.2 Automatic Registration 154
6.3.3 A Framework for SAR and Optical Data Registration in Case of HR Urban Images 155
6.3.3.1 Rigid Deformation Computation and Fourier--Mellin Invariant 155
6.3.3.2 Polynomial Deformation 157
6.3.3.3 Results 158
6.4 Fusion of SAR and Optical Data for Classification 158
6.4.1 State of the Art of Optical/SAR Fusion Methods 158
6.4.2 A Framework for Building Detection Based on the Fusion of Optical and SAR Features 161
6.4.2.1 Method Principle 161
6.4.2.2 Best Rectangular Shape Detection 162
6.4.2.3 Complex Shape Detection 163
6.4.2.4 Results 164
6.5 Joint Use of SAR Interferometry and Optical Data for 3D Reconstruction 165
6.5.1 Methodology 165
6.5.2 Extension to the Pixel Level 168
6.6 Conclusion 171
References 171
7 Estimation of Urban DSM from Mono-aspect InSAR Images 174
7.1 Introduction 174
7.2 Review of Existing Methods for Urban DSM Estimation 176
7.2.1 Shape from Shadow 177
7.2.2 Approximation of Roofs by Planar Surfaces 177
7.2.3 Stochastic Geometry 178
7.2.4 Height Estimation Based on Prior Segmentation 178
7.3 Image Quality Requirements for Accurate DSM Estimation 179
7.3.1 Spatial Resolution 179
7.3.2 Radiometric Resolution 181
7.4 DSM Estimation Based on a Markovian Framework 182
7.4.1 Available Data 182
7.4.2 Global Strategy 182
7.4.3 First Level Features 184
7.4.4 Fusion Method: Joint Optimization of Class and Height 185
7.4.4.1 Definition of the Region Graph 185
7.4.4.2 Fusion Model: MaximumA Posteriori Model 186
7.4.4.3 Optimization Algorithm 191
7.4.4.4 Results 191
7.4.5 Improvement Method 192
7.4.6 Evaluation 194
7.5 Conclusion 196
References 197
8 Building Reconstruction from Multi-aspect InSAR Data 199
8.1 Introduction 199
8.2 State-of-the-Art 200
8.2.1 Building Reconstruction Through Shadow Analysis from Multi-aspect SAR Data 200
8.2.2 Building Reconstruction from Multi-aspect Polarimetric SAR Data 201
8.2.3 Building Reconstruction from Multi-aspect InSAR Data 201
8.2.4 Iterative Building ReconstructionUsing Multi-aspect InSAR Data 202
8.3 Signature of Buildings in High-Resolution InSAR Data 202
8.3.1 Magnitude Signature of Buildings 203
8.3.2 Interferometric Phase Signature of Buildings 206
8.4 Building Reconstruction Approach 209
8.4.1 Approach Overview 209
8.4.2 Extraction of Building Features 211
8.4.2.1 Segmentation of Primitives 211
8.4.2.2 Extraction of Building Parameters 212
8.4.2.3 Filtering of Primitive Objects 213
8.4.2.4 Projection and Fusion of Primitives 214
8.4.3 Generation of Building Hypotheses 214
8.4.3.1 Building Footprint 215
8.4.3.2 Building Height 217
8.4.4 Post-processing of Building Hypotheses 218
8.4.4.1 Ambiguity of the Gable-Roofed Building Reconstruction 218
8.4.4.2 Correction of Oversized Footprints 221
8.5 Results 223
8.6 Conclusion 224
References 225
9 SAR Simulation of Urban Areas: Techniques and Applications 227
9.1 Introduction 227
9.2 Synthetic Aperture Radar Simulation Development and Classification 228
9.2.1 Development of the SAR Simulation 228
9.2.2 Classification of SAR Simulators 229
9.3 Techniques of SAR Simulation 231
9.3.1 Ray Tracing 231
9.3.2 Rasterization 231
9.3.3 Physical Models Used in Simulations 232
9.4 3D Models as Input Data for SAR Simulations 234
9.4.1 3D Models for SAR Simulation 234
9.4.2 Numerical and Geometrical Problems Concerning the 3D Models 234
9.5 Applications of SAR Simulations in Urban Areas 235
9.5.1 Analysis of the Complex Radar Backscattering of Buildings 235
9.5.2 SAR Data Acquisition Planning 237
9.5.3 SAR Image Geo-referencing 237
9.5.4 Training and Education 238
9.6 Conclusions 240
References 241
10 Urban Applications of Persistent Scatterer Interferometry 244
10.1 Introduction 244
10.2 PSI Advantages and Open Technical Issues 248
10.3 Urban Application Review 251
10.4 PSI Urban Applications: Validation Review 254
10.4.1 Results from a Major Validation Experiment 254
10.4.2 PSI Validation Results 255
10.5 Conclusions 256
References 257
11 Airborne Remote Sensing at Millimeter Wave Frequencies 260
11.1 Introduction 260
11.2 Boundary Conditions for Millimeter Wave SAR 261
11.2.1 Environmental Preconditions 261
11.2.1.1 Transmission Through the Clear Atmosphere 261
11.2.1.2 Attenuation Due to Rain 261
11.2.1.3 Propagation Through Snow, Fog, Haze and Clouds 261
11.2.1.4 Propagation Through Sand, Dust and Smoke 262
11.2.2 Advantages of Millimeter Wave Signal Processing 262
11.2.2.1 Roughness Related Advantages 262
11.2.2.2 Imaging Errors for Millimeter Wave SAR 263
11.3 The MEMPHIS Radar 264
11.3.1 The Radar System 264
11.3.2 SAR-System Configuration and Geometry 267
11.4 Millimeter Wave SAR Processing for MEMPHIS Data 268
11.4.1 Radial Focussing 268
11.4.2 Lateral Focussing 269
11.4.3 Imaging Errors 270
11.4.4 Millimeter Wave Polarimetry 273
11.4.5 Multiple Baseline Interferometry with MEMPHIS 275
11.4.6 Test Scenarios 277
11.4.7 Comparison of InSAR with LIDAR 279
References 281
Index 283