Night vision processing and understanding

This book systematically analyses the latest insights into night vision imaging processing and perceptual understanding as well as related theories and methods. The algorithm model and hardware system provided can be used as the reference basis for the general design, algorithm design and hardware design of photoelectric systems. Focusing on the differences in the imaging environment, target characteristics, and imaging methods, this book discusses multi-spectral and video data, and investigates a variety of information mining and perceptual understanding algorithms. It also assesses different processing methods for multiple types of scenes and targets. Taking into account the needs of scientists and technicians engaged in night vision optoelectronic imaging detection research, the book incorporates the latest international technical methods. The content fully reflects the technical significance and dynamics of the new field of night vision. The eight chapters cover topics including multispectral imaging, Hadamard transform spectrometry; dimensionality reduction, data mining, data analysis, feature classification, feature learning; computer vision, image understanding, target recognition, object detection and colorization algorithms, which reflect the main areas of research in artificial intelligence in night vision. The book enables readers to grasp the novelty and practicality of the field and to develop their ability to connect theory with real-world applications. It also provides the necessary foundation to allow them to conduct research in the field and adapt to new technological developments in the future. Foreword by Xiangqun Cui......Page 5 Preface......Page 9 Acknowledgements......Page 10 Contents......Page 11 1.1 Research Topics of Multidimensional Night-Vision Information Understanding......Page 15 1.1.1 Data Analysis and Feature Representation Learning......Page 16 1.1.2 Dimension Reduction Classification......Page 19 1.1.3 Information Mining......Page 22 1.2 Challenges to Multidimensional Night-Vision Data Mining......Page 24 References......Page 26 2.1 Multiplexing Measurement in Hyperspectral Imaging......Page 30 2.2.1 Traditional Denoising Theory of HTS......Page 32 2.2.2 Denoising Bound Analysis of HTS with S Matrix......Page 35 2.2.3 Denoising Bound Analysis of HTS with H Matrix......Page 38 2.3 Spatial Pixel-Multiplexing Coded Spectrometre......Page 40 2.3.1 Typical HTS System......Page 41 2.3.2 Spatial Pixel-Multiplexing Coded Spectrometre......Page 42 2.4 Deconvolution-Resolved Computational Spectrometre......Page 48 2.5 Summary......Page 54 References......Page 55 3.1 Infrared Image Super-Resolution via Transformed Self-similarity......Page 57 3.1.1 The Introduced Framework of Super-Resolution......Page 59 3.1.2 Experimental Results......Page 62 3.2 Hierarchical Superpixel Segmentation Model Based on Vision Data Structure Feature......Page 69 3.2.1 Hierarchical Superpixel Segmentation Model Based on the Histogram Differential Distance......Page 70 3.2.2 Experimental Results......Page 74 3.3 Structure-Based Saliency in Infrared Images......Page 82 3.3.1 The Framework of the Introduced Method......Page 83 3.3.2 Experimental Results......Page 89 3.4 Summary......Page 93 References......Page 94 4.1.1 New Adaptive Supervised Manifold Learning Algorithms......Page 98 4.1.2 Kernel Maximum Likelihood-Scaled LLE for Night-Vision Images......Page 100 4.2.1 Review of LDA and CMVM......Page 101 4.2.2 Introduction of the Algorithm......Page 103 4.2.3 Experiments......Page 105 4.3.1 Review of LPP......Page 109 4.3.2 Adaptive and Parameterless LPP (APLPP)......Page 110 4.3.3 Connections with LDA, LPP, CMVM and MMDA......Page 114 4.3.4 Experiments......Page 115 4.4.1 KML Similarity Metric......Page 120 4.4.2 KML Outlier-Probability-Scaled LLE (KLLE)......Page 123 4.4.3 Experiments......Page 124 4.4.4 Discussion......Page 131 4.5 Summary......Page 134 References......Page 135 5.1 Classification Methods......Page 137 5.1.1 Research on Classification via Semi-supervised Random Subspace Sparse Representation......Page 138 5.1.2 Research on Classification via Semi-supervised Multi-manifold Structure Regularisation (MMSR)......Page 139 5.2.1 Motivation......Page 140 5.2.2 SSM–RSSR......Page 142 5.2.3 Experiment......Page 146 5.3.1 Probability Semi-supervised Random Subspace Sparse Representation (P-RSSR)......Page 156 5.3.2 Experiment......Page 161 5.4.2 Multi-manifold Structure Regularisation (MMSR)......Page 169 5.4.3 Experiment......Page 174 5.5 Summary......Page 179 References......Page 181 6.1 Machine Learning in IM......Page 184 6.1.2 Feature Extraction and Classifier......Page 185 6.2.1 Denoising and Sparse Autoencoders......Page 186 6.2.2 LDAE......Page 188 6.2.3 Experimental Comparison......Page 191 6.3.1 Algorithm and Implementation of Detection System......Page 197 6.3.2 Experiments and Evaluation......Page 203 6.4 Summary......Page 206 References......Page 207 7.1.1 Investigation of Infrared Small-Target Detection......Page 209 7.1.2 Moving Object Detection Based on Non-learning......Page 210 7.1.3 Researches on Target Tracking Technology......Page 211 7.2.1 Framework of Object Detection......Page 212 7.2.2 Experimental Results......Page 214 7.3.1 Tracking Model Based on Global LARK Feature Matching and CAMSHIFT......Page 216 7.3.2 Target Tracking Algorithm Based on Local LARK Feature Statistical Matching......Page 219 7.3.3 Experiment and Analysis......Page 220 7.4 An SMSM Model for Human Action Detection......Page 225 7.4.1 Technical Details of the SMSM Model......Page 227 7.4.2 Experiments Analysis......Page 232 References......Page 240 8.1 Research on Colorization of Low-Light-Level Images......Page 243 8.2.1 Summary of the Principle of the Algorithm......Page 244 8.2.2 Mining of Multi-attribute Association Rules in Grayscale Images......Page 246 8.2.3 Colorization of Grayscale Images Based on Rule Mapping......Page 247 8.2.4 Analysis and Comparison of Experimental Results......Page 248 8.3 Multi-sparse Dictionary Colorization Algorithm Based on Feature Classification and Detail Enhancement......Page 254 8.3.1 Colorization Based on a Single Dictionary......Page 255 8.3.2 Multi-sparse Dictionary Colorization Algorithm for Night-Vision Images, Based on Feature Classification and Detail Enhancement......Page 256 8.3.3 Experiment and Analysis......Page 262 8.4 Summary......Page 271 References......Page 273