Cloud masks for optical EO data distingush cloudy pixels from cloud-free pixels. They may differentiate between serveral cloud types, i.e. opaque clouds and Cirrus clouds (that are transparent). Most land monitoring applications based on optical data require cloud-free images. Therefore, cloud masks are a product that is used early on in image processing for selecting suitable imagery for analysis (e.g. by screening images of an archive by the derived cloud cover percentage of the image). Therefore, cloud masks are made available as metadata by the EO data provider. Clouds are identified with threshoulding of reflectance values of the blue band and, to adapt for cloud/snow confusion, specific short-wave infrared (SWIR) bands.
Apply a cloud mask to optical EO data
New