Remote Sensing Intelligent Interpretation for Geology: From Perspective of Geological Exploration

✍ Scribed by Weitao Chen, Xianju Li, Xuwen Qin, Lizhe Wang

• Publisher: Springer
• Year: 2024
• Tongue: English
• Leaves: 240
• Category: Library

✦ Synopsis

This book presents the theories and methods for geology intelligent interpretation based on deep learning and remote sensing technologies. The main research subjects of this book include lithology and mineral abundance.

This book focuses on the following five aspects: 1. Construction of geology remote sensing datasets from multi-level (pixel-level, scene-level, semantic segmentation-level, prior knowledge-assisted, transfer learning dataset), which are the basis of geology interpretation based on deep learning. 2. Research on lithology scene classification based on deep learning, prior knowledge, and remote sensing. 3. Research on lithology semantic segmentation based on deep learning and remote sensing. 4. Research on lithology classification based on transfer learning and remote sensing. 5. Research on inversion of mineral abundance based on the sparse unmixing theory and hyperspectral remote sensing.

The book is intended for undergraduate and graduate students who are interested in geology, remote sensing, and artificial intelligence. It is also used as a reference book for scientific and technological personnel of geological exploration.

✦ Table of Contents

Preface
Contents
1 Geological Remote Sensing: An Overview
1.1 Description of Geological Remote Sensing
1.1.1 Concept and Principle of Geological Remote Sensing
1.1.2 Key Remote Sensing Techniques for Geological Research
1.1.3 Key Features for the Interpretation of Geological Remote Sensing
1.1.4 Technical System of Remote Sensing Technology for Geological Research
1.1.5 Applications of Geological Remote Sensing
1.2 Research Advance in Geological Remote Sensing
1.2.1 Payload: Developing from Single Optical Sensor to New Multimodal Sensors
1.2.2 Data Processing: Developing from “Rough” to “Precision”
1.2.3 Geological Application: Developing from Human–Computer Interaction to All-Factor Intelligent Interpretation
1.3 Intelligent Interpretation of Geological Remote Sensing
1.3.1 Technical System of Intelligent Interpretation of Remote Sensing for Geology
1.3.2 Advantages of Intelligent Interpretation of Geological Remote Sensing
1.4 Challenges and Future Directions of Geological Remote Sensing
1.4.1 Gaps and Challenges in Current State of Geological Remote Sensing
1.4.2 Research Directions of Geological Remote Sensing
References
2 Geological Remote Sensing Dataset Construction for Multi-level Tasks
2.1 Pixel-Level Dataset
2.1.1 Study Area
2.1.2 Data Sources
2.1.3 Data Preprocessing
2.1.4 Dataset Construction
2.2 Scene-Level Dataset
2.2.1 Study Area
2.2.2 Data Sources
2.2.3 Data Preprocessing
2.2.4 Dataset Construction
2.3 Semantic Segmentation-Level Dataset
2.3.1 Study Area
2.3.2 Data Sources
2.3.3 Data Preprocessing
2.3.4 Dataset Construction
2.4 Semantic Segmentation-Level Dataset Based on Multisource Data
2.4.1 Study Area
2.4.2 Data Sources
2.4.3 Dataset Construction
2.5 Prior Knowledge-Assisted Dataset
2.5.1 Study Area
2.5.2 Data Sources
2.5.3 Data Preprocessing
2.5.4 Dataset Construction
2.6 Transfer Learning Dataset
2.6.1 Study Area
2.6.2 Data Sources
2.6.3 Data Preprocessing
2.6.4 Dataset Construction
2.7 Transfer Learning Dataset for Prior Knowledge-Assisted Study
2.7.1 Study Area
2.7.2 Data Sources
2.7.3 Data Preprocessing
2.7.4 Dataset Construction
References
3 Lithological Classification Based on Large-Scale Pixel Neighborhood and VGGnet-Based Transfer Learning
3.1 Introduction
3.2 Methods
3.2.1 Construction of Model
3.2.2 VGG16 Convolutional Neural Network Model
3.2.3 VGG16 Transfer Learning Model
3.2.4 Accuracy Evaluation
3.3 Results
3.3.1 Experimental Environment and Setup
3.3.2 Full Image Prediction
3.3.3 Visual Evaluation of Prediction Results
3.3.4 Quantitative Accuracy Evaluation
3.4 Conclusion
References
4 Lithological Remote Sensing Scene Classification Based on Multi-view Data
4.1 Introduction
4.1.1 Research Background and Significance
4.1.2 Research Status
4.1.3 Research Objectives and Main Research Contents
4.2 Methods
4.2.1 Lithologic Scene Classification Based on Multi-view Remote Sensing Data Fusion
4.2.2 Accuracy Evaluation
4.3 Results and Discussion
4.3.1 Experimental Setup and Hyperparameter Optimization
4.3.2 Experimental Result
4.3.3 Discussions
4.4 Conclusion
References
5 Geological Lithology Semantic Segmentation Based on Deep Learning Method
5.1 Introduction
5.2 Methods
5.2.1 The Utilized Algorithms
5.2.2 Evaluation Metrics
5.3 Results
5.3.1 Experiment Setup
5.3.2 Model Performance
5.3.3 Visual Assessment
5.4 Discussions
5.5 Conclusion
References
6 Remote Sensing Lithology Intelligent Segmentation Based on Multi-source Data
6.1 Introduction
6.1.1 Research Background and Meaning
6.1.2 Research Status
6.1.3 Research Objectives and Research Content
6.2 Methods
6.2.1 Remote Sensing Lithology Semantic Segmentation Method Based on Adaptive Fusion of Multi-source Data
6.2.2 Remote Sensing Lithology Semantic Segmentation Method Based on Prior Knowledge
6.3 Results and Discussion
6.3.1 Evaluation of Test Set Accuracy of Multi-modal Data Adaptive Fusion Method
6.3.2 Test Set Accuracy Evaluation Using Methods that Incorporate Prior Knowledge
6.4 Conclusion
References
7 Prior Knowledge-Based Intelligent Model for Lithology Classification
7.1 Introduction
7.2 Methods
7.2.1 Lithological Scene Classification Based on Prior Knowledge and Improved Dense Connected Networks
7.2.2 Improved Dense Connectivity Network
7.2.3 Edge Enhancement
7.2.4 Experimental Setup and Environment
7.2.5 Evaluating Metrics
7.3 Results and Discussions
7.3.1 Comparative Experiment of 3EFFSA Model
7.3.2 Comparative Experiment of Prior-3EFFSA Model
7.3.3 Discussion
7.4 Conclusion
References
8 Multi-view Lithology Remote Sensing Scene Classification Based on Transfer Learning
8.1 Introduction
8.2 Methods
8.2.1 Lithologic Scene Classification Based on Transfer Learning
8.2.2 Lithologic Scene Classification Based on Multi-view Remote Sensing Data Fusion
8.2.3 Accuracy Evaluation
8.3 Results and Discussion
8.3.1 Experimental Setup and Hyperparameter Optimization
8.3.2 Experimental Result
8.4 Conclusion
References
9 Lithological Scene Classification Based on Model Migration and Fine-Tuning Strategy
9.1 Introduction
9.1.1 Overview of Transfer Learning
9.1.2 Deep Transfer Learning
9.2 Methods
9.2.1 Left Side as the Source Domain, and Right Side as the Target Domain
9.2.2 Right Side as the Source Domain, and Left Side as the Target Domain
9.3 Results and Analysis
9.3.1 Experimental Results and Analysis of Transfer Learning Model
9.4 Experimental Results and Analysis of Transfer Learning Based on Small Samples
9.5 Conclusion
References
10 Hyperspectral Remote Sensing Inversion of Mineral Abundance Based on Sparse Unmixing Method
10.1 Introduction
10.2 Methods
10.2.1 LMM
10.2.2 SUnSAL
10.2.3 SUnSAL-TV
10.2.4 MUA
10.3 Experimental Results and Discussion
10.3.1 Spectral Library
10.3.2 Simulation Datasets
10.3.3 Real Datasets
10.3.4 Evaluation Criterion
10.3.5 Experimental Analysis
10.4 Conclusion
References
11 Concluding Remarks
References