I'm using the
fmask algorithm to produce cloud masks for Sentinel2-L1C
.safe directories. The output is always a single
.img file. Afterwards I'd like to compute easy indices like the
ndvi on the area where there are no clouds according to the
fmask algorithm. As the L1C-data is still top of atmosphere data, I thought it'd be a good idea to first convert it to BOA with e.g. sen2cor.
So the workflow would look something like this:
L1C-Data --> Make cloudmask with
fmask --> mask entire
.safe-directory by outputted
<cloudmask>.img-file from fmask --> convert to to boa with sen2cor --> compute ndvi
Now I don't know if it's possible and even a good idea to assign a
NoData-value to certain areas of all bands within a L1C
.safe dataset. Maybe someone has an idea on how to this in a little smarter way...