I am using gdal_translate currently to cut a 60000x60000 geotiff into smaller 1000x1000 geotiff files. This works but it takes about 24 hours(and I have 44 of those big images to compute).
Im not sure if it would be faster but my theory is that using the gdal binding would be faster if it just loads the image into memory 1 time, then creates the new geotiff from that.(instead of running the gdal_translate command about 3600 times in series).
So the real question is, can I do that with the driver.createcopy command? Or do I have to use the driver.create method ? And would this be faster then actually starting gdal_translate 3600 times?
Currently I'm using gdal_translate like this from a batch file:
gdal_translate -of GTiff -srcwin 0 0 1000 1000 -co TFW=YES -co COMPRESS=JPEG -co PHOTOMETRIC=YCBCR -co JPEG_QUALITY=90 sourcefile.tif outfile.tif
I have seen that you can set the creation options like
PHOTOMETRIC=YCBCR with driver.createcopy but havent found any info on selecting a subwindow.