Hi,
I'm trying to find the location of a polygon over a cost-surface layer
(raster) that represents the minimum possible cost when you add the
pixels (cost) under it. The cost-surface has an irregular shape and
the size of the polygon (in this case a square) is about 1/1000 of the
size the cost layer. My approach (with a bash script) is to find
iteratively a random location for the polygon, rotate it at fixed
steps around it center (v.transform), calculate the sum of the cost
under it at each rotation (v.rast.stats), and keep the location that
produces the minimum sum of the cost. I could repeat the process, for
example 1.000.000 times, and keep the location that produces the
absolute minimum. I guess that the procedure should provide the right
answer (a location), but, is there any other way to do that in a more
efficient/elegant way? I'm also worry to be trapped around a local
minima.
Saludos, EKS
--
Eduardo Klein
Lab Sensores Remotos y
Centro de Biodiversidad Marina
Universidad Simón Bolívar
Caracas, Venezuela
ph (58) (212) 906-3111 ext 6700
fax (58) (212( 906-3111 ext 6701
_______________________________________________
grass-user mailing list
grass-user@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/grass-user
