USGSCSM and HRSC

[oleg-alexandrov]

I tried to run the HRSC example from here: https://github.com/USGS-Astrogeology/usgscsm/wiki/CSM-Stereo-Testing-Examples

knoten fails by a large margin when comparing the CSM and ISIS models. I used the same logic as here: #349.

The outputs are:

isis2csm_data
diff line diff sample
count 1.000000e+02 100.000000
mean -2.501009e+06 24778.453399
std 2.887109e+02 288.674276
min -2.501459e+06 24328.454250
25% -2.501259e+06 24528.454152
50% -2.501009e+06 24778.453422
75% -2.500759e+06 25028.452711
max -2.500559e+06 25228.452356

csm2isis_data
diff line diff sample
count 100.000000 100.000000
mean -920.490000 -2137.490000
std 288.675135 288.675135
min -1370.490000 -2587.490000
25% -1170.490000 -2387.490000
50% -920.490000 -2137.490000
75% -670.490000 -1887.490000
max -470.490000 -1687.490000

The differences are so huge that stereo fails.

[oleg-alexandrov]

I think I figured out the cause, which I mention at DOI-USGS/ale#424. I will close this as redundant.