The data is cut into 512 * 512 patches with overlapping
The data files are arranged as follows：
 LandSat8Biome
  |__512(Patch Size)
   |__B1:patch name.tif
   |__B2
   |__B3
   |__B4
   |__B5
   |__B6
   |__B7
   |__B8
   |__B9
   |__B10
   |__B11
   |__mask:patch name_mask.tif
   |__datatype_patch_name.csv(ContainingFile all patches' names, does not contain suffix)
  |__256(Patch Size)

This dataset comprises cloud masks for 513 1022-by-1022 pixel subscenes, at 20m resolution, sampled random from the 2018 Level-1C Sentinel-2 archive.
The data format is numpy array file（.npy）
The data is cut into 512 * 512 patches with overlapping
The data files are arranged as follows：
 Sentinel2
   |__NoShadow
   |__img:Patch Name.npy
   |__mask:Patch Name.npy
   |__datatype.csv(ContainingFile all patches' names, including suffix)

This data set simply outputs a pair of sentinel2 and landsat8 data in order to explore the cloud detection method across remote sensing platforms