[GRASS-user] Cost Distance Analysis

[Markus Metz]

LeeDaniel wrote:
>
> Hi all,
>
> Does anybody know how to perform a cost-distance analysis? I'm planning a
> street and have assigned various costs to different surfaces in my planning
> area. As I understand it, I need to perform the following steps:
>
> 1. r.cost using the starting point to generate a cost map for each pixel
> 2. r.walk to generate a raster with cumulative movement costs for each pixel
> from the point of origin
> 3. r.drain to generate the road
>
> My problem is, I don't want to model drainage; I want to find the route with
> the lowest costs possible to connect two points. Is there a way to calculate
> the path of least resistance from my origin point to a goal point? I've
> thought of using r.cost with two different points and then performing
> r.drain to connect the two. But would that work?

No, use either your origin point or your goal point as start point for
r.cost but not both points as start points, then use the other point
as start point for r.drain. You could also use either the origin point
or the goal point as start point for r.cost and the other point as
stop point for r.cost to speed up r.cost

It seems like r.drain
> produces paths that end at the point with the lowest cost, but that's not my
> goal - it gives me no control over where the street ends.

You do have control over where the street ends, if you use only one
start point for r.cost, because this will be the point with the lowest
cost. If I understand correctly, r.drain will then always stop at the
point used as start point for r.cost.

HTH,

Markus M