[Gdal-dev] Killing nodata images

[Frank Warmerdam]

Doug_Newcomb at fws.gov wrote:
> Hi folks,
>       I've written a python script to chop large .sid files into 10,000 x
> 10,000 pixel tiff , prior to conversion to .ecw format.  The sid files are
> countywide mosaics of aerial photographs.  Since not a lot of the counties
> in this state are square,  there can be large nodata sections in the sid
> files.  This means that some of the the blocks I chop out may be composed
> entirely of nodata areas.  I've been trying to figure out a quick way to
> determine if a file is entirely made up of nodata pixels while it's in tiff
> format, so I can delete it before wasting the cpu cycles  converting it to
> ecw.
> 
>        I'm fairly sure I'm missing something obvious due to  my lack of
> skill with python and lack of understanding  of how the gdal -python
> interface works, so all suggestions welcome.

Doug,

If you call GetMaximum() on a gdal.Band and it returns a very small value
then it is likely that the entire image is black or near black (assuming black
is your nodata value).  You could check if band.GetMinimum() ==
band.GetMaximum() == nodata value, but if you are coming from sid there may
be some noise that would make it inexact.  In that case checking for
min/max being close to the nodata value.

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 OSGF, http://osgeo.org