Even Rouault <even.rouault <at> mines-paris.org> writes:
> This is the very same topic that is discussed in > http://trac.osgeo.org/gdal/ticket/3947 . There's no solution to your problem, > but some background discussion that explains the current behaviour. This is close to a problem I had once with creating mosaics from individually warped orthophotos. Accurately calculated extents make pixels to slide a bit and individually warped images do not share any common canvas for their pixels. I sketched a plan for forcing the warped image to use a common canvas and found a python guy to make a program. I have been satisfied with the result. I believe that the same solution will work for you. You must widen the -te parameters that is calculated by crop_to_cutline to each direction so that they match exactly with some pixel row and line of the original image. See figures 3, 4, and 5 in http://www.scangis.org/scangis2007/papers/r3_rahkonen.pdf. The python code is there too. Without understanding anything about programming I suppose that the job is done with this: minmax = get_minmax(tm32_coords) minmax_wider = [ (int(math.floor(minmax[0][0])), int(math.floor(minmax[0][1]))), (int(math.ceil (minmax[1][0])), int(math.ceil (minmax[1][1]))), ] log("Extents (min,max): " + str(minmax), outfilename) log("Widened extents (min,max): " + str(minmax_wider), outfilename) -Jukka Rahkonen- _______________________________________________ gdal-dev mailing list [email protected] http://lists.osgeo.org/mailman/listinfo/gdal-dev
