SAR Image Analysis - A Computational Statistics Approach

Alejandro C. Frery, PhD, is Professor of Statistics and Data Science at the School of Mathematics and Statistics at Victoria University at Wellington, New Zealand. He earned his doctorate in Applied Computing at the National Institute for Space Research in Brazil.Jie Wu, PhD, is Associate Professor at the School of Computer Science, Shaanxi Normal University, China. He received his doctorate in Computer Science and Technology from Xidian University in China. Luis Gomez, PhD, is Associate Professor at the School of Telecommunications and Electronics Engineering, University of Las Palmas de Gran Canaria, Spain. He received his doctorate in Telecommunication Engineering from the Universidad de Las Palmas de Gran Canaria.

• Foreword xiiiPreface xviiAcknowledgments xxviiAcronyms xxxiIntroduction xxxiiiI.1 SAR xxxiiiI.2 Statistics for SAR xxxivI.3 The Book xxxvI.4 Commitment to Reproducibility and Replicability xxxix1 Data Acquisition 11.1 Introduction 11.2 SAR 31.2.1 The radar 41.2.2 What is SAR? 61.2.3 SAR systems 101.2.4 The synthetic antenna 161.3 Spatial resolution 201.4 SAR Imaging Techniques 231.5 The Return Signal: backscatter and speckle 281.5.1 Backscatter 281.5.2 Speckle 311.5.3 SAR geometric distortions 391.6 SAR Satellites 441.7 Preprocessing SAR data 531.8 Copernicus Open Access Hub 531.9 NASA Earth Data Open Data 561.10 Actual SAR Data Examples 571.10.1 Hawaii’s Big Island 571.10.2 Other examples 60Exercises 602 Elements of Data Analysis and Image Processing with R 732.1 Useful R Packages 732.1.1 Data loading 742.1.2 Data manipulation 762.2 Descriptive Statistics 782.2.1 Center tendency of data 782.2.2 Dispersion of data 812.2.3 Shape of data 842.3 Visualization 862.3.1 Rug and box plots 872.3.2 Histogram 882.3.3 Scattering Diagram 922.4 Statistics and Image Processing 942.4.1 Histogram based Image Transformation 942.4.2 Scattering based Analysis 982.5 The imagematrix package 1013 Intensity SAR Data and the Multiplicative Model 1053.1 The K distribution 1153.2 The G0 distribution 1173.3 The GH distribution 1213.4 Connection between Models 122Exercises 1234 Parameter Estimation 1274.1 Models 1284.1.1 The Bernoulli distribution 1284.1.2 The Binomial distribution 1284.1.3 The Negative Binomial distribution 1294.1.4 The Uniform distribution 1294.1.5 Beta distribution 1304.1.6 The Gaussian distribution 1314.1.7 Mixture of Gaussian distributions 1314.1.8 The (SAR) Gamma distribution 1324.1.9 The Reciprocal Gamma distribution 1324.1.10 The G0I distribution 1334.2 Inference by analogy 1344.2.1 The Uniform distribution 1344.2.2 The Gaussian distribution 1354.2.3 Mixture of Gaussian distributions 1354.2.4 The (SAR) Gamma distribution 1364.3 Inference by maximum likelihood 1364.3.1 The Uniform distribution 1374.3.2 The Gaussian distribution 1374.3.3 Mixture of Gaussian distributions 1384.3.4 The (SAR) Gamma distribution 1394.3.5 The G0 distribution 1404.4 Analogy vs. Maximum Likelihood 1414.5 Improvement by bootstrap 1424.6 Comparison of estimators 1434.7 An example 1444.8 The same example, revisited 1504.9 Another example 152Exercises 1575 Applications 1595.1 Statistical filters: Mean, Median, Lee 1605.1.1 Mean filter 1605.1.2 Median filter 1645.1.3 Lee filter 1675.2 Advanced filters: MAP and Nonlocal Means 1755.2.1 MAP Filters 1755.2.2 Nonlocal Means Filter 1775.2.3 Statistical NLM filters 1835.2.4 The statistical test 1895.3 Implementation Details 1915.4 Results 1935.5 Classification 1985.5.1 The image space of the SAR data 2055.5.2 The feature space 2075.5.3 Similarity criterion 2105.6 Supervised Image Classification of SAR Data 2125.6.1 The nearest neighbor classifier 2145.6.2 The K-nn method 2195.7 Maximum Likelihood Classifier 2235.8 Unsupervised Image Classification of SAR Data: The K-means classifier 2325.9 Assessment of Classification Results 236Exercises 2426 Advanced Topics 2496.1 Assessment of Despeckling Filters 2496.2 Standard Metrics 2496.2.1 Advanced Metrics for SAR Despeckling Assessment 2536.2.2 Completing the Assessment 2596.3 Robustness 2596.3.1 Robust inference 2606.3.2 The mean and the median 2616.3.3 Empirical Stylized Influence Function 2666.4 Rejoinder and Recommendations 2697 Reproducibility and Replicability 2737.1 What Is Reproducibility? 2737.2 What Is Replicability? 2747.3 Reproducibility and Replicability: Benefits for the Remote Sensing Community 2777.4 Recommendations for making “good science” 2787.5 Conclusions 283Index 301