Yes, I guess that what I meant too, you just formulate it better, in a
more general way. I find this interesting, as I have experimented with
weighted Voronoi polygons, but only in vector format. There is a
complete book on that methodolgy: Atsuyuki Okabe, Barry Boots, Kokochi
Sugihara, Sung Nok Chiu: Spatial Tesselations. Concepts and Applications
of Voronoi Diagrams. Wiley, Chichester 2000. It's really fascinating to
see how much you can do with this technology, not only in theory but
also in practice. Of course, the vector based algorithms are hard to
implement and can be very resource hungry. A lot of them are already
available however: just Google on "weighted voronoi". As always, a
raster based approach could be not only more efficient, but also
conceptually more general. There is an ArcGIS extension available that
does exactly this: see http://portal.acm.org/citation.cfm?id=1332465
With that in mind, if the algorithm you propose would be indeed an
approximation to weighted Voronoi polygons, *and* it wouldn't be all to
hard to implement (I have no idea about that), would it make sense to
propose this as a new RFC for GRASS?
Jan
Glynn Clements wrote:
Jan Hartmann wrote:
The problem with using r.cost is that you would need to know the cost
for each cell before you have created the polygons.
I think that the simplest accurate approach would be to modify
r.grow.distance.
Do you mean: adding a metric parameter to Euclidean, Squared, Manhattan,
and Maximum? Something like: compute on the basis of the value of
traversed cells?
I mean scale the distance by the value of the nearest non-null cell.
_______________________________________________
grass-user mailing list
grass-user@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/grass-user
