Heterogeneous SAR-optical fusion for near-real-time land use and land cover mapping under cloud contamination: A novel framework and global benchmark dataset
Researchers have developed CloudLULC-Net, a novel framework for land use and land cover mapping that effectively fuses Synthetic Aperture Radar (SAR) and optical remote sensing data. This method is designed to overcome the limitations of optical imagery frequently obscured by clouds and shadows. The framework incorporates techniques to suppress unreliable optical signals and adaptively aggregate information from both SAR and optical sources, creating a unified representation for LULC prediction. To support this work, a new benchmark dataset named CloudLULC-Set was created, containing over 40,000 SAR-optical-label triplets. AI
IMPACT This research offers a more robust approach to land cover mapping in challenging, cloud-prone regions, potentially improving environmental monitoring and resource management.