The prefilter options have changed. I also changed the text around so that it's little nicer.
…ample. I'm hoping to make calling map/map-PC.tif and nonmap/nonmap-PC.tif common options for the MOC and CTX example. This way when I want to run the tests, I don't have to keep looking at the Makefile.
If the user wants to use custom paths, they must define EXECDIR.
For a lot of them I changed correlation settings to use LOG with squared error. This is a lot faster than NCC. We don't need NCC for most of those because they're not too challenging and our subpixel seems to clean up most of our problems. I also added more comments about cost functions.
It's mostly parallel, so I'd recommend running make with a -j 3. The ortho projections don't match the DRG. I think this is because the triangulation between these two images is very poor. It would interesting to find out if a BA before all of this improved our results.
Have the sections actually go by their input and output filenames instead of easy-to-read names. This gives the benefit of running this makefile multiple times and not always have to keep remaking the map projected cubes files or stereo products. Also, the --xyz option is no longer valid.
…f the Stereo Pipeline Users Guide.