[R-sig-Geo] finding distance from a patch

[pgalpern]

I have not found a function in an existing package that will do this 
exactly, although there may be one.  However, here is an approach that 
will find the minimum distance from a point to a class based on 
functions in the raster package.  Note that this uses the grid distance 
and not the Euclidean distance for efficiency.  (Also, has not been 
thoroughly tested.)

DistancePointToClass <- function(classRaster, pointXY, focalClass) {
     rasterPointXY <- classRaster
     rasterPointXY[] <- NA
      rasterPointXY[cellFromXY(classRaster, pointXY)] <- 1
      distanceFrom <- gridDistance(rasterPointXY, origin = 1)
      minDistance <- min(distanceFrom[classRaster[] == focalClass])
      return(minDistance)
  }

The classRaster parameter can be your patches object.  If you need the 
minimum distance to any patch you can set all the patches to the same 
focalClass:

patches[!is.na(patches)] <- 1

Or if you need minimum distances to each patch, you can iterate through 
all the classes on patches using the focalClass parameter, although this 
will be very inefficient on large rasters.

You can find examples of usage of this function in this tutorial:  
http://nricaribou.cc.umanitoba.ca/R/RWorkshop3.pdf

-- 
Paul Galpern, PhD Candidate
Natural Resources Institute
70 Dysart Road
University of Manitoba
Winnipeg, Manitoba, Canada R3T 2M6
http://borealscape.ca



On 31/01/2012 5:38 PM, Lorenzo Cattarino wrote:
> Hi,
>
> I obtained the patches (i.e., clusters of connected cells - queen's case) from a raster object. What I am trying to do is to calculate the euclidean distance between (the border of) a patch and a cell in the raster (outside the patch). What I am looking for is the minimum euclidean distance between the cell outside the patch and the patch.
>
> I would like to find the coordinates of the points of each patch to calculate the distance between each of those points and the cell of interest. Then I will choose the smallest value.
>
> require(raster)
> landscape<- matrix (sample(c(1,0), 100, T), 10)
> w<- raster(landscape)
> #plot(w)
> patches<- clump(w)
> #windows()
> #plot(patches)
>
> Is there a function that might help doing that?
>
> thanks
> Lorenzo
> _______________________________________________
> R-sig-Geo mailing list
> R-sig-Geo at r-project.org
> https://stat.ethz.ch/mailman/listinfo/r-sig-geo