ASTER files are provided via FTP as HDF format files with accompanying XML files containing metadata, including footprint an which instruments are "on" for that image.
We are primarily interested in the VNIR data layers, which are HDF subdatasets 1, 2, and 3. These are all the same pixel dimensions, and each subdataset is an 8-bit grey-scale image corresponding to a visible ro near-infrared wavelength.
Thus, the three layers can be combined into a single RGB image for a "false color" final product. The subdatases are also georeferenced, so the resulting image can be a GeoTiff with three separate color bands.
The approach is to use the GDAL utilities gdal_translate to extract the 3 subdatasets, then gdal_merge.py to merge them into a single GeoTiff, and finally gdal2tiles.py to create a tiled KML for viewing in Google Earth. From there the resulting KML might be integrated into a master KML library via network links with TimeSpan tags.
Problem: gdal_translate does not seem to extract a georeference that gdal_merge.py can use. Might neet to run it through gdalwarp first. Need to test this more thoroughly
Idea: might take the source code for gdal_merge.py and gdal2tiles.py and modify to do everything in one step using the GDAL api