| 摘要 | 第3-5页 |
| Abstract | 第5-7页 |
| Abbreviations | 第22-24页 |
| Chapter 1 Introduction | 第24-33页 |
| 1.1 Motivation | 第24-26页 |
| 1.2 Fundamental Concepts | 第26-29页 |
| 1.3 Thesis Contributions | 第29-30页 |
| 1.4 Thesis Organization | 第30-33页 |
| Chapter 2 Prior Works for Feature Extratcion and Classification of SAR Imagery | 第33-46页 |
| 2.1 Hand-designed features based approaches | 第33-39页 |
| 2.1.1 Bayesian Compressive Sensing (BCS) with Scattering Centers Features Algorithm | 第33-34页 |
| 2.1.2 A Wavelet Decomposition Method | 第34-35页 |
| 2.1.3 SAR-ATR Using Discriminative Graphical Models | 第35-36页 |
| 2.1.4 SAR-ATR Using Multi-layer Neural Network (ANN) | 第36-37页 |
| 2.1.5 SAR-ATR Using SIFT Method | 第37-38页 |
| 2.1.6 Texture and Geometric Features Combination | 第38-39页 |
| 2.2 Deep Convolutional Neural Networks based approaches | 第39-46页 |
| 2.2.1 A-Conv Nets | 第40-41页 |
| 2.2.2 DCNN Based on Auto-encoder | 第41-42页 |
| 2.2.3 DCNN with Data Augmentation Techniques | 第42-44页 |
| 2.2.4 SAR-ATR using DCNN and SVMs Combination | 第44-46页 |
| Chapter 3 SAR Image Target Classification Method Using Hand-designed Features and Support Vector Machines | 第46-77页 |
| 3.1 Introduction | 第46-47页 |
| 3.2 SAR Oriented Visual Saliency Model and Directed Acyclic Graph Support Vector Metric Based Target Classification | 第47-60页 |
| 3.2.1 SAR Oriented GBVS Model | 第48-52页 |
| 3.2.2 Gabor Feature Extraction | 第52-53页 |
| 3.2.3 HOG Feature Extraction | 第53-54页 |
| 3.2.4 Feature Fusion Using DCA | 第54-57页 |
| 3.2.5 DAG-SVML Classification Using the Mahalanobis Distance Metric | 第57-60页 |
| 3.3 Experimental Results and Analysis | 第60-76页 |
| 3.3.1 Datasets | 第60-62页 |
| 3.3.2 Experimental Results | 第62-63页 |
| 3.3.3 Results on MSTAR Public Mixed Target Dataset | 第63-66页 |
| 3.3.4 Results on MSTAR Public T-72 Variants Dataset | 第66-68页 |
| 3.3.5 Results on MSTAR with Large Depression Angle Variations | 第68-69页 |
| 3.3.6 Results on Outlier Rejection | 第69-70页 |
| 3.3.7 Detection Results | 第70-73页 |
| 3.3.8 Comparison with Different Fusion Methods | 第73-74页 |
| 3.3.9 Comparison with the State-of-the-art Classifiers | 第74页 |
| 3.3.10 Comparison with Recent Representative Methods | 第74-76页 |
| 3.4 Conclusion | 第76-77页 |
| Chapter 4 Bag-of-Visual-Words Based Feature Extraction and Selection for SAR Target Classification | 第77-93页 |
| 4.1 Introduction | 第77-78页 |
| 4.2 Bo VW Based Feature Extraction and Classification | 第78-85页 |
| 4.2.1 Bo VW Feature Representation | 第79-80页 |
| 4.2.2 Classification | 第80-81页 |
| 4.2.3 Experimental Results | 第81-85页 |
| 4.3 An Efficient Feature Selection for SAR Target Classification | 第85-92页 |
| 4.3.1 Feature Selection | 第86-88页 |
| 4.3.2 Evaluation and Results | 第88-92页 |
| 4.4 Conclusion | 第92-93页 |
| Chapter 5 Deep Feature Extraction and Combination for Synthetic Aperture Radar Target Classification | 第93-111页 |
| 5.1 Introduction | 第93-94页 |
| 5.2 Framework of the Classification Method | 第94-101页 |
| 5.2.1 Data Pre-Processing | 第95-96页 |
| 5.2.2 Feature Extraction | 第96-97页 |
| 5.2.3 Feature Fusion | 第97-99页 |
| 5.2.4 Learn the Mahalanobis Distance Metric and K-NN Classification | 第99-101页 |
| 5.3 The Proposed VGG-S1 Architecture | 第101-103页 |
| 5.3.1 Local Response Normalization | 第102页 |
| 5.3.2 Dropout Regularization | 第102-103页 |
| 5.3.3 SVM Classifier | 第103页 |
| 5.4 Experimental Results and Analysis | 第103-110页 |
| 5.4.1 Implementation details | 第103-106页 |
| 5.4.2 Results on MSTAR Public Mixed Target Dataset | 第106-107页 |
| 5.4.3 Results on MSTAR Public T-72 Variants Dataset | 第107-109页 |
| 5.4.4 Results on MSTAR with Large Depression Angle Variations | 第109-110页 |
| 5.5 Conclusion | 第110-111页 |
| Chapter 6 Very Deep Multi-Canonical Correlation Analysis for SAR Image Target Classification | 第111-128页 |
| 6.1 Introduction | 第111-113页 |
| 6.2 Proposed SAR Target Classification Method | 第113-120页 |
| 6.2.1 SAR Oriented Network | 第113-117页 |
| 6.2.2 MCCA Based Feature level Fusion | 第117-119页 |
| 6.2.3 Support Vector Machine (L2-SVM) | 第119-120页 |
| 6.3 Experimental Results and Analysis | 第120-127页 |
| 6.3.1 Implementation details | 第121页 |
| 6.3.2 Results on MSTAR Public Mixed Target Dataset | 第121-123页 |
| 6.3.3 Results on MSTAR Public T-72 Variants Dataset | 第123-124页 |
| 6.3.4 Results on MSTAR with Large Depression Angle Variations | 第124-125页 |
| 6.3.5 Comparison with Different Fusion Methods | 第125-126页 |
| 6.3.6 Comparison with Recent Representative Methods | 第126-127页 |
| 6.4 Conclusion | 第127-128页 |
| Conclusions | 第128-131页 |
| References | 第131-142页 |
| List of Publications | 第142-145页 |
| Acknowledgements | 第145-146页 |
| Resume | 第146页 |
