* This dataset supports online use.Click here to jump.

Clouds in satellite images pose a significant challenge for downstream applications. A major problem facing current cloud removal research is the lack of comprehensive benchmarks and sufficiently large and diverse training datasets. To address this problem, a research team from Cornell University and Columbia University launched AllClear in 2024, the largest public cloud removal dataset, containing 23,742 globally distributed regions of interest (ROIs), covering a variety of land use patterns, and a total of 4 million images. The related paper results are "AllClear: A Comprehensive Dataset and Benchmark for Cloud Removal in Satellite Imagery", which has been accepted by NeurIPS.

Each ROI includes a full time series capture for the full year 2022, including:

• multispectral optical images from Sentinel-2 and Landsat 8/9;
• Synthetic Aperture Radar (SAR) images from Sentinel-1;
• Ancillary remote sensing products such as cloud masks and land cover maps.