On 21/02/2012 22:36, Frank Warmerdam wrote:
On Tue, Feb 21, 2012 at 1:17 PM, Armin Burger<armin.burger@xxxxxxx> wrote:
Then I tried to set the NODATA value to 0 using Gdal-Python, and afterwards
they had

Band 1 Block=3000x1 Type=Byte, ColorInterp=Red
NoData Value=0
Band 2 Block=3000x1 Type=Byte, ColorInterp=Green
NoData Value=0
Band 3 Block=3000x1 Type=Byte, ColorInterp=Blue
NoData Value=0

But the result was the same, when using "-r average" a 1-2 pixel area of the
original NODATA values along the border to the DATA areas seem to have got
pixels with higher values than 0 (at least in Mapserver they cannot any more
be "hidden" with OFFSITE). Is there a simple way to write out the overviews
to separate Tiff files so that I can check them directly?

Armin,

You can use the gdal/apps/dumpoverviews app to write out overviews. It may
not be built and installed by default. If you confirm the overviews are built
wrong then please file a ticket, and provide with it a small sample file with
which you encounter the problem.

Best regards,

Frank

I checked the overviews after dumping them into Tiff files. I could not see anything strange with them along the border DATA-NODATA. The supposed NODATA pixels all had value 0 in all bands. So the problem seems a bit related how Mapserver handles the files. Strangely enough that the problem does not appear when the OV are created with nearest resampling. Being somehow more related to Mapserver I filed a ticket there

http://trac.osgeo.org/mapserver/ticket/4210

But the ticket owner are already you anyway ;-)

I could not upload a small test image due to file size restrictions (the zip has 3 MB). Should I send the file to you or someone else via mail?

I discovered an additional strange effect when using the layer processing parameter
PROCESSING "RESAMPLE=BILINEAR"

which will create partially colored effects along the DATA-NODATA line. All of this happens just in the case that this line is not exactly horizontal or vertical but oblique due to image warping. I will create another ticket for this. So far it seems that I can use no "smoothing" via resampling, neither with overviews nor with the layer definition without having some "border effects".

Thanks

Armin

_______________________________________________
gdal-dev mailing list
gdal-dev@xxxxxxxxxxxxxxx
http://lists.osgeo.org/mailman/listinfo/gdal-dev