[Gdal-dev] Problem with gdalwarp -tps for large files

[Frank Warmerdam]

Jan Hartmann wrote:
> There seems to be a problem with gdalwarp -tps with control points for 
> rasters > 10000*10000 pixels. The problem appears right at the 
> beginning, when gdalwarp tries to determine the size of the output 
> raster. It does not appear with smaller rasters, and also not  when the 
> output file already exists, in other words when its size is already 
> known. It looks like an overflow problem to me. I could file a bug 
> report, but the test-file is quite large (278M).  Any idea?

Jan,

I'd appreciate your filing a bug with the particulars and with a
gdalinfo report of the input file.  I can dummy up a similarly sized
file at my end to try and reproduce the problem.

I have had problems with the -tps algorithm "flipping out" on me from
time to time and have been assuming it was some sort of numerical
instability but it is possible there is a more conventional sort of
error at the root of the issue.

Best regards,
-- 
---------------------------------------+--------------------------------------
I set the clouds in motion - turn up   | Frank Warmerdam, warmerdam at pobox.com
light and sound - activate the windows | http://pobox.com/~warmerdam
and watch the world go round - Rush    | President OSGeo, http://osgeo.org