Deep Learning for Hyperspectral Image Analysis and Classification (Engineering Applications of Computational Methods, 5)

✍ Scribed by Linmi Tao, Atif Mughees

• Publisher: Springer
• Year: 2021
• Tongue: English
• Leaves: 217
• Category: Library

✦ Synopsis

This book focuses on deep learning-based methods for hyperspectral image (HSI) analysis. Unsupervised spectral-spatial adaptive band-noise factor-based formulation is devised for HSI noise detection and band categorization. The method to characterize the bands along with the noise estimation of HSIs will benefit subsequent remote sensing techniques significantly.

This book develops on two fronts: On the one hand, it is aimed at domain professionals who want to have an updated overview of how hyperspectral acquisition techniques can combine with deep learning architectures to solve specific tasks in different application fields. On the other hand, the authors want to target the machine learning and computer vision experts by giving them a picture of how deep learning technologies are applied to hyperspectral data from a multidisciplinary perspective. The presence of these two viewpoints and the inclusion of application fields of remote sensing by deep learning are theoriginal contributions of this review, which also highlights some potentialities and critical issues related to the observed development trends.

✦ Table of Contents

Preface
Acknowledgements
Contents
1 Introduction
1.1 Applications of Hyperspectral Images
1.2 Challenges in Hyperspectral Image Classification
1.3 Research Objective
1.4 Research Achievements
1.4.1 Noise Reduction/Band Categorization of HSI
1.4.2 Unsupervised Hyperspectral Image Segmentation
1.4.3 Deep Learning Based HSI Classification Techniques
1.5 Organization of the Book
References
2 Hyperspectral Image and Classification Approaches
2.1 Introduction to Hyperspectral Imaging
2.1.1 Hyperspectral Imaging System
2.1.2 Why Hyperspectral Remote Sensing
2.2 Review of Machine Learning Based Approaches for Hyperspectral …
2.2.1 Hyperspectral Image Interpretation Taxonomy
2.3 Hyperspectral Remote Sensing Image Dataset Description
2.3.1 Indian Pine: AVIRIS Dataset
2.3.2 Pavia University: ROSIS Dataset
2.3.3 Houston Image: AVIRIS Dataset
2.3.4 Salinas Valley: AVIRIS Dataset
2.3.5 Moffett Image: AVIRIS Dataset
2.3.6 Washington DC Mall Hyperspectral Dataset
2.4 Classification Evaluation Measures
2.5 Literature Review
2.5.1 HSI Noise/Redundancy Detection
2.5.2 Deep Learning Based Algorithms
References
3 Unsupervised Hyperspectral Image Noise Reduction and Band Categorization
3.1 Proposed Methodology
3.1.1 Preprocessing Toward Initial Segmentation
3.1.2 Cluster-Size Factor
3.1.3 Cluster-Shift Factor
3.1.4 Cluster Spatial-Spectral Contextual Difference Factor
3.1.5 Band-Noise Factor (BNF)
3.1.6 The HSI Process and Noise Model
3.1.7 Noise Estimation
3.2 Experimental Results
3.2.1 HSI Datasets
3.2.2 Synthetic Dataset
3.2.3 Experiments on Real HS Data
3.2.4 Discussion of rET Parameters
3.2.5 Discussion of Weight-Subfactor Parameters
3.2.6 Noise-Level Estimation
3.2.7 HSI Classification
3.3 Summary of the Proposed Unsupervised Hyperspectral Image Noise Reduction and Band Categorization Method
References
4 Hyperspectral Image Spatial Feature Extraction via Segmentation
4.1 Proposed Methodology
4.1.1 Preprocessing Toward Initial Segmentation
4.1.2 Boundary Model
4.1.3 Channel–Group Merge Criteria
4.2 Experimental Approach and Analysis
4.2.1 Hyperspectral Datasets
4.2.2 Adjustment of Weight Factors
4.2.3 Grouping and Merging Methods
4.2.4 Experimental Results and Comparison
4.2.5 Evaluation Measures
4.3 Summary of the Proposed Hyperspectral Image Spatial Feature Extraction via Segmentation Method
References
5 Integrating Spectral-Spatial Information for Deep Learning Based HSI Classification
5.1 Preprocessing
5.1.1 Band Selection
5.1.2 Hyper-Segmentation-Based Spatial Feature Extraction
5.2 SAE Based Shape-Adaptive Deep Learning for Hyperspectral …
5.2.1 Feature Extraction
5.2.2 Experimental Results and Performance Comparisons
5.2.3 Summary of the Proposed Integration of Spectral-Spatial Information Method for Deep Learning Based HSI Classification
5.3 DBN-Based Shape-Adaptive Deep Learning for Hyperspectral Image Classification
5.4 Hyper-Segmentation Based DBN for HSI Classification
5.4.1 Extraction of Spectral-Spatial Information of Spatial Segments via DBN
5.4.2 Experimental Results and Performance Comparisons
5.4.3 Summary of the Proposed DBN-Based Shape-Adaptive Deep Learning Method for Hyperspectral Image Classification
5.5 PCANet-Based Boundary-Adaptive Deep Learning for Hyperspectral Image Classification
5.5.1 SANet-Based Spectral-Spatial Classification Network
5.5.2 Experimental Analysis and Performance Comparisons
5.5.3 Parameter Setting
5.5.4 Summary of the Proposed PCANet-Based Boundary-Adaptive Deep Learning Method for Hyperspectral Image Classification
5.6 Summary of the Proposed Deep Learning Based Methods for Hyperspectral Image Classification
References
6 Multi-Deep Net Based Hyperspectral Image Classification
6.1 Multi-Deep Belief Network-Based Spectral–Spatial …
6.1.1 Spectral-Adaptive Segmented DBN for HSI Classification
6.1.2 Spectral–Spatial Feature Extraction by Segmented DBN
6.1.3 Experimental Results and Performance Comparisons
6.1.4 Experimental Setup
6.1.5 Spectral–Spatial HSI Classification
6.1.6 Summary of the Proposed Multi-Deep Net-Based Hyperspectral Image Classification Method
6.2 Hyperspectral Image Classification Based on Deep Auto-Encoder …
6.3 Proposed Methodology
6.3.1 HMRF-EM Segmentation
6.3.2 Final Segmentation with Preserved Edges
6.3.3 SAE Pixel-Wise Classification
6.3.4 Majority Voting
6.4 Experimental Results and Performance Comparisons
6.5 Summary of the Proposed Hyperspectral Image Classification Method Based on Deep Auto-encoder and Hidden Markov Random Field
6.6 Hyperspectral Image Classification Based on Hyper-segmentation and Deep Belief Network
6.6.1 Proposed Methodology
6.6.2 Experimental Results and Performance Comparison
6.7 Summary of the Proposed Deep Learning-Based Methods for Hyperspectral Image Classification
References
7 Sparse-Based Hyperspectral Data Classification
7.1 Introduction
7.2 Related Methods
7.3 Proposed Approach
7.3.1 Sparse Representation for Hyperspectral Data Using a Few Labeled Samples
7.3.2 Homotopy
7.3.3 Sparse Ensemble Framework
7.4 Experimental Results and Comparison
7.4.1 Effect of Parameter Selection on Classification Accuracy
7.4.2 AVIRIS Hyperspectral Image
7.4.3 Washington DC Mall Image
7.4.4 Kennedy Space Center and Salina A Hyperspectral Datasets
7.4.5 Time Comparison of General LP Sparse and Homotopy-Based Sparse Representations
7.5 Discussion
7.5.1 Sparsity of Computed Solution
7.5.2 Sparse Classification
7.5.3 Sparse Solution Analysis
7.5.4 Sparse Reconstruction
7.6 Summary
References
8 Challenges and Future Prospects
8.1 Future Prospects