Hi Greg, Il giorno Sun, 31 Jan 2010 14:58:19 -0500 Greg Coats <gregco...@mac.com> ha scritto:
> Regarding the diff function > "diff": computes the difference between 2 raster bands (b1 - b2) > Since the result of diff could be a negative number, for example > 99-150=-51, I suggest that there be an additional function that > provides the absolute value of diff, which would be 51 in this > example. Greg Of course, it would not be a problem to include but my original idea was to provide only base functions and use them to get complex expressions. In this case, what you need can be obtained using two stacked virtual files: - diff (using appropriate band data type and SourceTransferType) - mod Is it a reasonable approach for you? Even more interesting for common tasks should be to compute abs((b1-b2)/(b1+b2)) > On Jan 31, 2010, at 12:18 PM, Antonio Valentino wrote: > > > Hi Matt, > > > > Il giorno Mon, 25 Jan 2010 20:57:58 -0600 > > Matt Klaric <m...@klaric.com> ha scritto: > > > >> On Mon, Jan 25, 2010 at 8:38 PM, Frank Warmerdam > >> <warmer...@pobox.com>wrote: > >> > >>> I don't think the VRT approach supports actually adding > >>> different images together, only overlaying and rescaling. > >>> > >>> > >> However, pixel functions could be used in a VRT file to accomplish > >> the desired outcome. > >> > >> --matt > > > > Indeed, it would require a pixel function for sum and the ScaleRatio > > property of ComplexSource. > > > > I have just submitted a patch > > > > http://trac.osgeo.org/gdal/ticket/3367 > > > > that provides a set of standard pixel functions, including sum (of 2 > > or more raster bands). > > > > Using this set of functions it should be quite easy to write a > > python program that performs generic bands arithmetic. best regards -- Antonio Valentino _______________________________________________ gdal-dev mailing list gdal-dev@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/gdal-dev
