View article

Clouds obstruct the viewing of the earth's surface by satellites operating in the visual and infrared spectrum. On the basis of their particular spectral behaviour, they can be detected, and eliminated from the image before further processing. Thus, a reliable cloud mask is an essential and early step in the processing chain of VEGETATION imagery.

The objective of the project is to develop such a cloud mask, as well as a snow mask and a cloud shadow mask. A'good'cloud mask depends on a variety of factors, which have all been included, when the multicriteria decision was made to proceed with the current masks, of which the methodology is described in the next paragraph. Parameters that describe the quality of a cloud mask include: flexibility, compatibility (with the current cloud masking algorithm), speed, interpretability (or acceptability to the user community), results. Five broad types of algorithmic approaches have been developed, and many subtypes of these have been tested in an initial phase.