Mary Jo Brodzik wrote:
Dear gdal-dev list members,
I'm using a python tool called the regionator
(http://code.google.com/p/regionator/wiki/ Welcome) on a large
(14352x7440) png image. The regionator uses the GDAL library to
resample subsets of the original image into much smaller chunk images.
These images are set up at increasing levels of detail in a kml file
that make the download of this information into a googleEarth client
much more efficient.
The png is color and includes transparency; its approximate resolution
is 500 meters. I basically have 4 colors in the original image, black,
white, blue and yellow. The output images that regionator is producing
are great, but when I zoom in to distances where the original pixel
sizes are visible, it looks like the resampling did some averaging
across the 4 colors, so I'm getting blue sections surrounded by a blue
halo that bleeds into white, and yellow and blue values bleeding into
each other, etc.
I'd like to control the resampling, and force it to do nearest-neighbor
rather than averaging to eliminate the halos. It looks to me like the
regionator is doing the resampling with calls to ReadRaster and
WriteRaster in in a class called extractor.py, but I don't know enough
about the GDAL libraries to know how to change them. Is it possible to
force the resampling to be nearest neighbor? extractor.py isn't
terribly large or complicated, so I'm including it here (below).
Mary Jo,
ReadRaster() cannot itself be modified to do bilinear resampling.
I would suggest you try out gdal2tiles.py which does something similar
to regionator but with many more options.
http://www.gdal.org/gdal2tiles.html
You might also find MapTiler of interest:
http://www.maptiler.org/
Best regards,
--
---------------------------------------+--------------------------------------
I set the clouds in motion - turn up | Frank Warmerdam, [email protected]
light and sound - activate the windows | http://pobox.com/~warmerdam
and watch the world go round - Rush | Geospatial Programmer for Rent
_______________________________________________
gdal-dev mailing list
[email protected]
http://lists.osgeo.org/mailman/listinfo/gdal-dev
