[GRASS-stats] r.texture and MASK

[Agustin Lobo]

Could be the presence of a MASK causing r.texture to complain
"Negative or no data pixel found..."?
This is not described in the doc... and our raster map
does not have any null value.

Anyway, this is not the way it should be in grass, as
the MASK is supposed to prevent any program to
process the null part of the raster. Am I wrong?
I think that there is no other solution (again) than modifying
the code so that those elements in the null part of the MASK
are not counted for the co-occurrence matrix.  Setting null
cells in the MASK as 0 in the raster is not a good option, as
windows close to the border of the non-0 values will be
including 0 and thus producing spurious results.

Has anyone done any work in this direction?

Agus
-------------- next part --------------
A non-text attachment was scrubbed...
Name: alobolistas.vcf
Type: text/x-vcard
Size: 251 bytes
Desc: not available
Url : http://lists.osgeo.org/pipermail/grass-stats/attachments/20090417/f0baeba5/alobolistas.vcf