You should use gdalwarp only, no ned for gdal_merge.py.

The order of input datasets is important, so in your case you should use
gdalwarp south.tif north.tif  merge.tif

Plus you don't need a mask, the last input file will override the first one.

Etienne

On Fri, Apr 13, 2012 at 4:09 PM, Gregory, Matthew
<[email protected]> wrote:
> Hi all,
>
> I don't think I'm understanding correctly how to use masking bands.  My goal 
> is to create a mosaic between two adjacent rasters using mask bands on each 
> raster to define the valid data.  As a test, I have two single band GeoTiffs 
> (north.tif and south.tif) that overlap by about 400 rows.  I use the 
> following Python script to create a small mask at the top center of the south 
> image which creates the desired .msk file:
>
>  from osgeo import gdal, gdalconst
>  ds = gdal.Open('south.tif', gdalconst.GA_ReadOnly)
>  ds.CreateMaskBand(2)
>  rb = ds.GetRasterBand(1)
>  mask = rb.GetMaskBand()
>  data = mask.ReadAsArray()
>  data[:] = 255
>  data[0:500,2500:3000] = 0
>  mask.WriteArray(data)
>  ds = None
>
> Then I try to merge the two images together:
>
>  gdal_merge.py -o merge.tif -of GTiff -createonly north.tif south.tif
>  gdalwarp north.tif south.tif merge.tif
>
> I was expecting that the masked pixels from south.tif would be considered 
> NODATA allowing the north.tif pixels to 'shine through' in that location.  
> Instead the original pixel values from south.tif also show up in the mosaic.
>
> thanks, matt
> _______________________________________________
> gdal-dev mailing list
> [email protected]
> http://lists.osgeo.org/mailman/listinfo/gdal-dev
_______________________________________________
gdal-dev mailing list
[email protected]
http://lists.osgeo.org/mailman/listinfo/gdal-dev

Reply via email to