This book introduces remotely sensed image processing for urban areas using optical and synthetic aperture radar (SAR) data and assists students, researchers, and remote sensing practitioners who are interested in land cover mapping using such data. There are many introductory and advanced books on optical and SAR remote sensing image processing, but most of them do not serve as good practical guides. However, this book is designed as a practical guide and a hands-on workbook, where users can explore data and methods to improve their land cover mapping skills for urban areas. Although there are many freely available earth observation data, the focus is on land cover mapping using Sentinel-1 C-band SAR and Sentinel-2 data. All remotely sensed image processing and classification procedures are based on open-source software applications such QGIS and R as well as cloud-based platforms such as Google Earth Engine (GEE).
The book is organized into six chapters. Chapter 1 introduces geospatial machine learning, and Chapter 2 covers exploratory image analysis and transformation. Chapters 3 and 4 focus on mapping urban land cover using multi-seasonal Sentinel-2 imagery and multi-seasonal Sentinel-1 imagery, respectively. Chapter 5 discusses mapping urban land cover using multi-seasonal Sentinel-1 and Sentinel-2 imagery as well as other derived data such as spectral and texture indices. Chapter 6 concludes the book with land cover classification accuracy assessment.
Author(s): Courage Kamusoko
Series: Springer Geography
Publisher: Springer
Year: 2021
Language: English
Pages: 130
City: Cham
Preface
How is this Workbook Organized?
Conventions Used in this Workbook
Data and Online Resources
Contents
1 Geospatial Machine Learning in Urban Environments: Challenges and Prospects
Abstract
1.1 Introduction
1.1.1 Background
1.1.2 Cloud and Desktop Open Source Software Applications
1.2 Study Area and Data
1.2.1 Harare Metropolitan Area
1.2.2 Satellite Imagery
1.2.2.1 Sentinel-1 Data
1.2.2.2 Sentinel-2 Data
1.3 Preliminary Image Processing in GEE
1.3.1 Lab 1. Processing Sentinel-2 Imagery
1.3.2 Lab 2. Exploring Sentinel-1 Imagery
1.3.3 Lab 2(a). Accessing Sentinel-1 Imagery
1.3.4 Lab 2(b). Examining Sentinel-1 Orbit Properties
1.3.5 Lab 2(c). Examining Sentinel-1 Polarization
1.3.6 Lab 2(d). Exporting Sentinel-1 Imagery
1.4 Summary
1.5 Additional Exercises
References
2 Exploratory Analysis and Transformation for Remotely Sensed Imagery
Abstract
2.1 Introduction
2.1.1 Preparing Training Areas
2.1.2 Land Cover Classification Scheme
2.1.3 Spectral and Texture Indices
2.2 Exploratory Data Analysis and Image Transformation
2.2.1 Lab 1. Preparing Training Data
2.2.2 Lab 2. Creating Spectral Plots
2.2.3 Lab 3. Computing Spectral Indices using Sentinel-2 Imagery
2.2.4 Lab 4. Computing Texture Indices using Sentinel-2 Imagery
2.2.5 Lab 5. Computing Texture Indices using Sentinel-1 Imagery
2.3 Summary
2.4 Additional Exercises
References
3 Mapping Urban Land Cover Using Multi-seasonal Sentinel-2 Imagery, Spectral and Texture Indices
Abstract
3.1 Introduction
3.1.1 Background
3.1.2 Land Cover Mapping Using Multi-seasonal Imagery and Other Derived Data
3.2 Land Cover Mapping Labs
3.2.1 Lab 1. Mapping Land Cover Using Multi-seasonal Sentinel-2 Imagery
3.2.2 Lab 2. Mapping Land Cover Using Multi-seasonal Sentinel-2 Imagery and Other Derived Data
3.3 Summary
3.4 Additional Exercises
References
4 Mapping Urban Land Cover Using Multi-Seasonal Sentinel-1 Imagery and Texture Indices
Abstract
4.1 Introduction
4.1.1 Background
4.1.2 Synthetic Aperture Radar (SAR) Basics
4.2 Land Cover Mapping Labs
4.2.1 Lab 1. Mapping Land Cover Using Multi-Seasonal Sentinel-1 Imagery
4.2.2 Lab 2. Mapping Land Cover using Multi-seasonal Sentinel-1 Imagery and Texture Indices
4.3 Summary
4.4 Additional Exercises
References
5 Improving Urban Land Cover Mapping
Abstract
5.1 Background
5.2 Land Cover Mapping Labs
5.2.1 Lab 1. Mapping Land Cover Using Multi-Seasonal Sentinel-1 and Sentinel-2 Imagery
5.2.2 Lab 2. Mapping Land Cover Using Multi-Seasonal Sentinel-1 and Sentinel-2 Imagery and Other Derived Data
5.3 Summary
5.4 Additional Exercises
References
6 Land Cover Classification Accuracy Assessment
Abstract
6.1 Background
6.1.1 Sampling Design
6.1.2 (a) Stratified Random Sampling
6.1.3 (b) Determine Sample Size
6.1.4 Response Design
6.1.5 (a) Spatial Assessment Unit
6.1.6 (b) Sources of Reference Data
6.1.7 (c) Reference Labeling Protocol
6.1.8 (d) Defining Agreement
6.1.9 Analysis
6.1.10 (a) The Confusion (Error) Matrix
6.1.11 (b) Estimating Accuracy
6.1.12 (c) Estimating Area
6.2 Performing Accuracy Assessment
6.2.1 Lab 1. Sample Design
6.2.2 Lab 2. Response Design
6.3 Summary
6.4 Additional Exercises
References
Appendix_1
A.1. Additional Learning Resources
