[Gdal-dev] Different projection results between gdalwarp and gdal python bindings

[Jose Luis Gomez Dans]

Frank,
> > Jose Luis Gomez Dans wrote:
> 
> > > I am finding some inconsistencies between the gdalwarp utility and the
> [...] 
> > I've tried reproducing your problem with FWTools 1.3.4 on linux but the
> > results match exactly.
> 
> Mine is self-compiled, but due to dependencies and so on, I might have
> ended up with different sets of GDAL-related stuff on my machine. I will have
> a deep inspection to see whether this is the case or not. 

After checking whether everything was in order, I decided to install FWTools (1.3.5, the latest and greatest). The problem is still there. I have also tried the same on a clean FWTools  installation on win32 (it's an old install I have around, with GDAL 1.3.2), and I get the same results I reported yesterday. 

Converting a Shapefile results in no problems, but converting this GeoTIFF files is problematic. Given this, I guess that the CSV files with projection parameters are OK (otherwise, ogr2ogr reprojections should fail).

I also notice that the pixel size is slightly different (see my original post), so could the problem stem from estimating where the data are grid or corner referenced? Wild guessing on my side here :) 

I have put a sample GeoTIFF file in <http://simanctel.110mb.com/tmp/MOD15A2.A2007217.h17v05.005.2007230234148.tif>, if anyone wants to give it a go (it's only 119kb).

Many thanks for your help,
Jose



-- 
Psssst! Schon vom neuen GMX MultiMessenger gehört?
Der kanns mit allen: http://www.gmx.net/de/go/multimessenger