[R-sig-Geo] Convert SpatialPointsDataFrame to a SpatialGridDataFrame

[Roger Bivand]

On Wed, 23 Sep 2009, Brendan Malone wrote:

> Hi everyone,
>
>
>
> I would like to know a good way in which to convert a
> SpatialPointsDataFrame to a SpatialGridDataFrame.
>
> The reason that I ask is that when I change the CRS of a
> SpatialGridDataFrame, I inadvertently create a point data frame with the
> new CRS.

Your input data are in geographical coordinates, and you are projecting 
(no datum transformation) to UTM. This inevitably means that while the 
input points are evenly spaced, the projected points are not, and do not 
constitute a regular grid (see any introduction to projections). To get a 
regular grid out, you will have to warp, that is interpolate, to a regular 
grid from the irregular points that you have generated.

Either use an interpolator such as idw() in gstat (see the task view for 
more), or use for example gdal_warp externally.

This is getting to be a FAQ, isn't it? Could someone put it in the R-wiki 
under spatial data?

Roger

>
>
>
> To better illustrate my problem:
>
>
>
> I have a grid file
>
>
>
> grid90m <- readGDAL("edgeroi_dem.asc")
>
> names(grid90m) <- "edgeroi_dem"
>
>
>
> I give it a CRS
>
>
>
> proj4string(grid90m)<- CRS("+init=epsg:4283")
>
>
>
> Then I change it to a projected CRS
>
>
>
> grid90.UTM <- spTransform(grid90m, CRS("+init=epsg:32755"))
>
>
>
> Get a summary
>
> summary(grid.UTM)
>
> Object of class SpatialPointsDataFrame
>
> Coordinates:
>
>      min       max
>
> x  738742  790718.3
>
> y 6643860 6679923.8
>
> Is projected: TRUE
>
> proj4string :
>
> [+init=epsg:32755 +proj=utm +zone=55 +south +ellps=WGS84 +datum=WGS84
> +units=m +no_defs +towgs84=0,0,0]
>
> Number of points: 220805
>
> Data attributes:
>
>   Min. 1st Qu.  Median    Mean 3rd Qu.    Max.
>
> 180.2   200.7   225.0   254.1   280.6   988.6
>
>
>
> I have tried various ways to convert this points file to grid via a few
> methods with resultant failure. Essentially I need to input some measure
> of the pixel size.
>
> My original thoughts were that the problem was to do with the
> asymmetrical shape of my DEM layer. This may be true, but it would be
> useful if people could provide some ideas on how to get around this.
>
>
>
> Cheers,
>
>
>
> Brendan Malone
>
> PhD candidate
>
> University of Sydney
>
>
>
>
>
>
> 	[[alternative HTML version deleted]]
>
> _______________________________________________
> R-sig-Geo mailing list
> R-sig-Geo at stat.math.ethz.ch
> https://stat.ethz.ch/mailman/listinfo/r-sig-geo
>

-- 
Roger Bivand
Economic Geography Section, Department of Economics, Norwegian School of
Economics and Business Administration, Helleveien 30, N-5045 Bergen,
Norway. voice: +47 55 95 93 55; fax +47 55 95 95 43
e-mail: Roger.Bivand at nhh.no