Paper dataset: Using High-Resolution Satellite Imagery and Deep Learning to Map Artisanal Mining Spatial Extent in the Democratic Republic of the Congo

This dataset accompanies the following publication: Using High-Resolution Satellite Imagery and Deep Learning to Map Artisanal Mining Spatial Extent in the Democratic Republic of the Congo. Remote Sensing, 17(24), 4057. https://doi.org/10.3390/rs17244057 Overview This dataset contains paired satellite image tiles and binary segmentation masks for Artisanal and Small-scale Mining (ASM) sites in the Democratic Republic of the Congo (DRC). It is intended to support training and evaluation of deep learning models for ASM mapping using optical and SAR imagery. Below you can see the structure of the dataset. dataset/├── ps/ # PlanetScope tiles (optical, 4-band)│ ├── images/ # RGB + NIR GeoTIFF tiles│ └── masks/ # Binary ASM masks (0 = background, 1 = ASM)└── s1/ # Sentinel-1 tiles (SAR, 2-band) ├── images/ # VV and VH polarization GeoTIFF tiles └── masks/ # Binary ASM masks (0 = background, 1 = ASM) Data Specifications PlanetScope Sentinel-1 Source Planet-NICFI ESA Copernicus Bands Blue, Green, Red, NIR VV, VH GSD 4.77 m 10 m Tile size 256 × 256 px 128 × 128 px Number of tiles 782 782 Format GeoTIFF GeoTIFF The PlanetScope and Sentinel-1 tiles are spatially overlapping — corresponding tiles in ps/ and s1/ cover the same geographic area in the Eastern Democratic Republic of the Congo. Masks are binary rasters at the same resolution and spatial extent as the corresponding image tiles: 0 = Non-ASM (background) 1 = Artisanal and Small-scale Mining (ASM) Ground truth labels were derived from field survey data collected by [e.g. IPIS] and processed as described in the accompanying paper. License This dataset contains data from two sources with different licenses: PlanetScope tiles (ps/ directory): derived from Planet-NICFI basemaps. Imagery © Planet Labs Inc. All use subject to the Planet NICFI Participant License Agreement. Use of these tiles is restricted to non-commercial purposes in support of the NICFI program's goals (tropical forest monitoring, climate change, biodiversity conservation, and sustainable development). Sentinel-1 tiles and all masks (s1/ directory and all masks/ subdirectories): Licensed under Creative Commons Attribution 4.0 International (CC BY 4.0). Citation If you use this dataset, please cite: Pasanisi et al. (2025). Using High-Resolution Satellite Imagery and Deep Learning to Map Artisanal Mining Spatial Extent in the Democratic Republic of the Congo. Remote Sensing, 17(24), 4057. https://doi.org/10.3390/rs17244057