Image Pre-processing, Feature Generation and Classification in Remote Sensing
NFDI4EarthXTUD
About This Course
This self-paced course consists of 11 Jupyter Notebooks and provides learners with the full processing chain for remote sensing classification tasks. It includes four modules: raster data handling and pre-processing, data enhancement and feature extraction, unsupervised classification, and supervised classification with accuracy assessment. Students will work with real-world Sentinel-2 imagery, apply machine learning techniques, and evaluate classification accuracy (Access this course on GitLab).
Level
Intermediate, Advanced
Requirements
Basic Knowledge of Python Programming
Subject Area
Remote Sensing, Environmental Data Science, Geoinformatics, Machine Learning
What You Will Learn
- Python Basics and Math Functions
- Data Handling with Pandas
- Global Temperature Time Series Analysis
- Raster Data with Xarray and Sentinel-2
- Feature Extraction using Principal Component Analysis
- Texture Feature Extraction with GLCM
- Edge and Texture Enhancement using Gabor Filters
- Unsupervised Classification with KMeans
- Hierarchical Clustering and Visualization
- Supervised Classification using Random Forest
- Support Vector Machines and Accuracy Assessment
Resources
EduPilot "Image pre-processing, feature generation and classification in remote sensing" by Christine Wessollek, Matthias Forkel, Luisa Schmidt