Incorrect matches plotted when multiple .match files present

[dshean]

My directory contains two .match files, as --num-matches-from-disparity was included during the parallel_stereo call. L__R.match is written during stereo_pprc, and a second match file (long filename) is written during stereo_tri. Current stereo.py just uses first filename from glob list.

Should revisit assumptions about L_sub and orthos and match files. The L_sub might not be orthorectified, but could be in original image coords (lines/samples) rather than map coordiantes. The L__R match file should always be consistent with L_sub/R_sub. Should test using the mapproject stereo workflow and the "original" stereo workflow without running mapproject first.