distance between raster cell centroids and spatial points

Dear list,

I want to calculate the distance between the centroid of each cell in a
raster ?ras? and the closest village (villages are stored as
SpatialPointsDataFrame).

ras is as follow:
     class       : RasterLayer
     dimensions  : 5027, 3386, 17021422  (nrow, ncol, ncell)
     resolution  : 27.28532, 27.28532  (x, y)


I can obtain what I want with a sub-sample of ras that has less cells, and
here are my codes:

*# sub sample ras*
*sub_ras <- aggregate(ras, fact=20, fun=mean)*
     class       : RasterLayer
     dimensions  : 252, 170, 42840  (nrow, ncol, ncell)
     resolution  : 545.7063, 545.7063  (x, y)

*# extract x and y coord from each cell of the raster*
*centroid <- xyFromCell(sub_ras, c(1:ncell(sub_ras)))  *

*# combine x and y into a spatial point object*
*centroid <- SpatialPoints(cbind(centroid[,1], centroid[,2]))*

*# calculate distance of each centroid to the closest village and then turn
it into a raster of *
*# extent ?sub_ras' and values 'dist'*
*dist <- nncross(as(centroid,"ppp"), as(villages, "ppp"))$dist  *

*rasterdist2village <- rasterize(centroid, sub_ras, dist)*
*plot(rasterdist2village)*


I then tried to run the same codes using all cells of ras. After 15 minutes
the ?rasterize? function was still running and so I aborted the procedure.

I wonder whether there is a more elegant and computationally faster way of
doing what explained above?

Any input is highly appreciated

Best,
Gabriele






--
Gabriele Cozzi
Postdoctoral Research Associate
Population Ecology Research Group
http://www.popecol.org

Zurich University
Institute of Evolutionary Biology and Environmental Studies
Winterthurerstr. 190
8057 Zurich - Switzerland
E-mail: gabriele.cozzi at uzh.ch
Phone: +41(0)44 635 49 12
Fax: +41(0)16355711
http://www.ieu.uzh.ch

        [[alternative HTML version deleted]]