[GRASS-user] walking path network generation

[Jason Jorgenson]

Hi everyone.  I have about 100 points on a normal size ASTER DEM.  I want to
find the optimal walking paths connecting them all.  I guess I want to build
a road network, but based on walking time.  I have generated cost surfaces
for each point using r.walk, and have then used r.drain to find paths from
each site to all other sites, using the corresponding cost surface generated
from the origin point.
I then thinned each of the resultant path rasters.  Once that was completed,
I converted each to a vector.  Finally, I used v.patch to create a
horrifying black tangle of chaos that I don't believe can be untangled.

So, the question is, where did I go (a) right and (b) wrong?  I also didn't
like the way that r.drain seems to not follow the same route when going to
nearby points using the r.walk cost surface.

I read through the email exchanges between M Barton and others wrt r.walk,
and understand that r.walk should not use slope as a friction map, as it
already accounts for that intrinsically.  It is in this way that I used
r.walk (friction map was all 1's) .  Any advice?

Hope this email finds all who read it well :-)

Jason



Jason Jorgenson
Ph.D. Candidate
University of Liverpool

"Never let the fact that nothing is really known about past events stand in
your way"
 --Paul Bahn, regarding archaeology
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.osgeo.org/pipermail/grass-user/attachments/20090718/00292382/attachment.html