A co-registered Sentinel-1 (SAR) + Sentinel-2 (optical) flood extent mapping dataset,
built on STURM-Flood, for training and evaluating deep learning fusion models for
flood segmentation.
STURM-Fusion extends the optical-only STURM-Flood dataset by pairing each tile with
temporally matched Sentinel-1 SAR imagery, enabling direct comparison of SAR-only,
optical-only, and multi-modal fusion segmentation models on identical flood events.
Paper: MSc dissertation —… See the full description on the dataset page:
https://huggingface.co/datasets/taylor-millin-reade/STURM-fusion-24.