Am 29.06.2010 18:23, schrieb Chaitanya kumar CH:
Stefan,

I haven't tested the script but doesn't the current code copy the alpha band? If it does, it might be more preferable to copy all the pixels and the alpha band.
Yes the current code copies the alpha band, but it uses the same nodata settings as the other bands. This gets interesting if I want to merge overlapping images which define their valid area by an alpha channel. In the current code, I would have to specify nodata=0 to get the alpha channel to merge correctly. This is counterintuitive, as I expect images with alpha to merge correctly without specifying a value for nodata. If nodata isn't specified the merged alpha of overlapping images is incorrect (as opaque areas get overwritten with the transparent parts (nodata) )

Please create a ticket at http://trac.osgeo.org/gdal/newticket and attach your patch.
Done... http://trac.osgeo.org/gdal/ticket/3669

Regards
Stefan

On Tue, Jun 29, 2010 at 8:19 PM, Stefan Klug <[email protected] <mailto:[email protected]>> wrote:

    Hi,

    gdal_merge.py doesn't handle tiff files with alpha masks so I've
    added this functionality.

    My patch changes the behaviour of gdal_merge as follows:
    - If a band is recognized as Alpha Band it is automatically
    treated as having a NODATA value of 0.
    I'm not sure if this is the way to go. But it makes some sense, as
    the resulting alpha channel is a correct alpha merge, independent
    of the nodata value used for the other bands.
    - If a band has mask flag GMF_ALPHA set, the mask band controls
    the parts that get written to the destination

    Would be great if someone could review and possibly checkin the patch.

    Regards
       Stefan




    _______________________________________________
    gdal-dev mailing list
    [email protected] <mailto:[email protected]>
    http://lists.osgeo.org/mailman/listinfo/gdal-dev




--
Best regards,
Chaitanya kumar CH.
/tʃaɪθənjə/ /kʊmɑr/
+91-9494447584
17.2416N 80.1426E


_______________________________________________
gdal-dev mailing list
[email protected]
http://lists.osgeo.org/mailman/listinfo/gdal-dev

Reply via email to