Loading...
Loading...
Available on 1 platform
Sign in to view source links and access this dataset
513 1022-by-1022 pixel subscenes sampled from the 2018 Level-1C Sentinel-2 archive provide pixel-wise cloud masks. The dataset was annotated semi-automatically using the IRIS toolkit by Alistair Francis and John Mrziglod from University College London. It includes binary classification tags for each subscene covering 11 surface types and 7 cloud types.
For training cloud shadow masks specifically, users are advised to remove the 89 images where shadows were not annotatable, as this subset contains the most difficult examples.