Adapt Algorithm depending on LC type

[Interne CNES]

WASP currently produces poor results in at least two conditions:

• High cloud coverage (tropical forests, equatorial regions,...)
• Snow/Ice coverage (polar regions, mountains)

The goal shall be to adapt WASP within reasonable conditions to produce higher quality L3A products over those areas. Land Cover (LC) maps might help determining these problematic areas in the first place.

Some ideas that already yielded better results on some sites:

• Add condition to update LandPixel only if Red > 0 && Green > 0 as MAJA often tends to overcorrect cloud/terrain shadows and set reflectances to 0.
• Add condition to not update Pixels when Blue > Threshold where Threshold=some value. As such, WASP will basically do its own cloud detection on L2A products, which may or may not yield better results depending on the region.