Hi,
From a classification process we are developing, we created a
unsupervised classification from the image. We then want to select
random locations from the unsup raster as a training area for the 2nd
level of classification. I think it can be done by: r.to.vect,
v.random, v.extract. The problem is r.to.vect reports an out of
memory when converting the whole region. I tried r.neighbors to to
reduce 2-3 pixels clumps. Probable solution would be r.to.vect at a
smaller region then v.overlay. Is there a more straighforward way?
Like filtering the raster to include only large clumps from a randomly
selected points before r.to.vect?
cheers,
maning
--
|---------|----------------------------------------------------------|
| __.-._ |"Ohhh. Great warrior. Wars not make one great." -Yoda |
| '-._"7' |"Freedom is still the most radical idea of all" -N.Branden|
| /'.-c |Linux registered user #402901, http://counter.li.org/ |
| | /T |http://esambale.wikispaces.com|
| _)_/LI
|---------|----------------------------------------------------------|