[GRASSLIST:2583] Was: Re: Moire effect in aspect map

[Christof Bigler]

Looking at the map in ArcGIS, it says that it has 10418 cols and 8949 
rows, the same number as gdalinfo shows.
I "solved" the problem with the shifted grid cells and the missing row 
and column by importing a *.e00 raster file, exported from ArcGIS.
r.info gives now consistent boundaries, number of columns/rows and 
resolution. However, the problem with the inconsistency between grass 
and gdal still remains, at least maybe for other Grass users.

The second and original problem with the moire effect in the aspect map 
also remains, even 1) when I derive aspect from the newly imported 
*.e00 raster map, or 2) when I import the aspect map from a *.e00 file, 
which has been derived in ArcGIS, or 3) when I export the map as TIFF. 
BUT, there is no such effect, when I look e.g. at the map 'aspect' from 
the spearfish data set, which has exactly the same resolution as the 
default settings.

Any comments on this issue?

Christof

On 12.02.2004, at 09:53, John Gillette wrote:

> OK. There's a problem here.  If you look at the coordinates from grass
> versus gdal, Grass has shifted the NW coordinate 1/2 resolution size
> to the north and west. This would be consistent if grass was 
> interpreting
> the gdal coordinates to be the centers of the pixel cells.  BUT, then
> grass ends up with a SE coordinate that is 1.5 cells to the north and 
> west
> of the gdal SE corner and therefore ends up with 1 fewer row and 
> column than
> the gdal data.  (I would have expected this to be .5 cells to the SE, 
> if grass
> was using cell centers and I would have expected 1 MORE row and 
> column.)
> This is based on the information you show below.
>
> gdal, on the other hand, appears to be using the coordinates as cell 
> edges
> because the number of columns it reports is (E-W)/resolution.  If it 
> was
> using the coordinates as cell centers, the number would be 
> [(E-W)/resolution]+1.
>
> It would be interesting to confirm exactly how many rows and columns 
> are
> actually in the data.  Perhaps you could use gdal_translate to 
> translate the
> file into tiff and use some picture program to tell you what the actual
> dimensions of the tiff are. (Also look for your moire pattern.)
>
> In any event, there seems to be an inconsistency here between grass 
> and gdal.
>
> John