This project's aim was to develop deep learning models to map artisanal and small-scale mining (ASM) sites using high-resolution satellite imagery and optical-SAR. The models were based on the U-Net architecture and were trained on a dataset of labeled images of ASM sites and non-ASM sites (binary ground truth).
This project stemmed out from my MSc thesis in Geo-Information Science at Wageningen University & Research (GRS lab). The thesis was supervised by dr. Robert Masolele and dr. Johannes Reiche.
- Python 3.10+
- Poetry (dependency management tool used in this project)
- Clone the repository:
git clone https://github.com/yourusername/asm-mapping.git cd asm-mapping - Install dependencies with poetry:
or, depending on the Poetry version you installed, run the command below and then paste in the command line its output:
poetry install
poetry source env activate source /home/your/username/asm-mapping/.venv/bin/activate
- Activate the virtual environment (not necessary if you used the 'source' command above):
poetry shell