I'm trying to find if GRASS includes some-sort of way to detect/eliminatd cloud cover and shadows from LANDSAT Images.
The only thing I found was this: http://svn.osgeo.org/grass/grass-addons/imagery/i.landsat.acca/description.html
But:
1- has anyone tried/used this? Is this useful and working?
2- can I use it to detect-and-eliminate clouds?
Thanks
Antonio
__________ Information from ESET NOD32 Antivirus, version of virus signature database 5486 (20100928) __________
Yes, I've used it in the past using 6.3 svn. I highly recommend it.
You need to convert your Landsat data to top-of-atmosphere radiance or
top-of-atmosphere reflectance first before running i.landsat.acca: http://grass.osgeo.org/wiki/GRASS_AddOns#i.landsat.toar
Off-topic: I hope the landsat tools (i.topo.corr, i.landsat.toar,
i.landsat.acca, i.landsat.dehaze, i.attcorr) gets into the core in the
future.
It is currently under active development. Useful - hope so
2- can I use it to detect-and-eliminate clouds?
Should be the idea. The best way is to contact the author who
is very responsive (see manual page).
Cheers
Markus
Thanks
Antonio
__________ Information from ESET NOD32 Antivirus, version of virus signature
database 5486 (20100928) __________
--
cheers,
maning
------------------------------------------------------
"Freedom is still the most radical idea of all" -N.Branden
wiki: http://esambale.wikispaces.com/
blog: http://epsg4253.wordpress.com/
------------------------------------------------------
Off-topic: I hope the landsat tools (i.topo.corr,
i.landsat.toar, i.landsat.acca, i.landsat.dehaze, i.attcorr) gets into
the core in the future.
yes, we're working on .toar and .acca now; I think .dehaze needs some minor
cleanup & pythonization for trunk but no one is working on that right now;
and i.atcorr is already there in all branches.
(today's .acca got a little messy in svn, give it a few days to get cleaned
back up)
you'll want to read the PDF paper by Irish in the refs section of the .acca
man page, it discusses where the cloud detection algorithm does well and
where it does not do well: for opaque clouds it does well, for thin semi-
transparent clouds it does not do well. creating a cloud MASK from the
output of i.landsat.acca with r.reclass or r.mapcalc is very simple.
