Hi Don:

The GRASS GIS Module v.net.path will compute shortest path distances between points associated with a vector line dataset. The short story here is that GRASS is a separate open-source software package that integrates very closely with R (e.g, spatial data move
transparently between the two systems)

You might check out the GIS links on the attached page to get a bird's-eye view of GRASS:

http://nceas.ucsb.edu/scicomp/SoftwareLinks.html

You would have to be creative in your use of coordinate systems, but you could solve this
problem using GRASS GIS.

Regards,
Rick Reeves
Don MacQueen wrote:
This question is basically off topic, since it isn't truly an R question, though it is certainly related to r-sig-geo. I ask here in the hopes that someone can suggest a direction for me to look.

I have a what is essentially a floor plan of a building (lines representing walls, breaks in the lines representing doors, and so on). I would like to have an algorithm that can calculate the shortest distance between any two points (actually, large numbers of pairs of points) that does not intersect any line (pass through any wall).

My real dream is to use these distances instead of Euclidean distances as input to a gaussian random fields algorithm such as those in the RandomFields package.

I've done some web-searching for libraries of geometric algorithms that might include this task, and have come across references to Dystra's algorithm, and some mention of line of sight calculations, as well as some algorithms for robotics (how the robot finds a path from here to there, passing around obstacles represented as polygons). So I have some leads.

None the less, I'd really appreciate any suggestions from the folks here on R-sig-geo.

Thanks
-Don
begin:vcard
fn:Rick  Reeves
n:Reeves;Rick 
org:UC Santa Barbara;National Center For Ecological Analysis and Synthesis
adr:Suite 300;;735 State Street;Santa Barbara;CA;93101;USA
email;internet:[EMAIL PROTECTED]
title:Rick Reeves,  Scientific Programmer/Analyst
tel;work:805 892 2533
url:http://www.nceas.ucsb.edu
version:2.1
end:vcard

_______________________________________________
R-sig-Geo mailing list
R-sig-Geo@stat.math.ethz.ch
https://stat.ethz.ch/mailman/listinfo/r-sig-geo

Reply via email to