FireSR: A Dataset for Super-Resolution and Segmentation of Burned Areas

# FireSR Dataset ## Overview FireSR is a dataset designed for the super-resolution and segmentation of burned areas due to wildfires. It includes data for all wildfire events in Canada from 2017 to 2023 that are larger than 2000 hectares. The dataset is intended to support high-resolution daily monitoring and effective wildfire management using machine learning techniques. ## Dataset Structure The dataset is organized into the following directories: - **S2**: Contains Sentinel-2 images. - **pre**: Pre-fire Sentinel-2 images. - **post**: Post-fire Sentinel-2 images. - **mask**: Contains National Burned Area Composite (NBAC) polygons, which are the ground truth masks for the burned areas. - **MODIS**: Contains post-fire MODIS images. Each directory contains GeoTIFF (.tif) files named according to the format: `CA_<year>_<province>_<id>.tif`, where: - `CA` stands for Canada. - `<year>` is the year of the wildfire event. - `<province>` is the province code (e.g., AB for Alberta, BC for British Columbia). - `<id>` is a unique identifier for the wildfire event. ## File Structure FireSR/ │ ├── dataset/ │ ├── S2/ │ │ ├── post/ │ │ │ ├── CA_2017_AB_204.tif │ │ │ ├── CA_2017_AB_2418.tif │ │ │ └── ... │ │ ├── pre/ │ │ │ ├── CA_2017_AB_204.tif │ │ │ ├── CA_2017_AB_2418.tif │ │ │ └── ... │ ├── mask/ │ │ ├── CA_2017_AB_204.tif │ │ ├── CA_2017_AB_2418.tif │ │ └── ... │ ├── modis/ │ ├── CA_2017_AB_204.tif │ ├── CA_2017_AB_2418.tif │ └── ... ## License This dataset is licensed under the Creative Commons Attribution 4.0 International License (CC BY 4.0). You are free to share and adapt the material as long as appropriate credit is given. ## Contact For any questions or further information, please contact: - Name: Eric Brune - Email: [e-mail]