Seeking guidance for SfS using lunar South Pole images.

7 views
Skip to first unread message

Albert Kim

unread,
Mar 12, 2026, 1:48:49 PMMar 12
to Ames Stereo Pipeline Support
Hello all,

Recently, I have been attempting to use LRONAC south pole images for SfS. I have figured out an alignment pipeline that works pretty well, resulting in a DTM that looks fairly accurate in valid areas. I am using --save-dem-with-nodata to invalidate any areas that are in shadow in all input images, which results in many small holes (usually at the bottom of craters) in the final DTM, as well as some big ones. I understand that it is an expected behavior, and am looking for ways to fill these holes. Things I have tried:
  • sfs_blend with the base (LOLA) DTM. This worked as expected, but has its limitations due to the low resolution.
  • curvature, borderline data, etc. constraints. Works all right for smaller holes but creates artifacts in other areas. I would also prefer to have many holes and fill some of them afterwards, instead of having no holes and many subtle inaccuracies.
  • various methods of interpolation, to little success.
Any advice or guidance would be appreciated.

Best,
Albert

Oleg Alexandrov

unread,
Mar 12, 2026, 1:55:23 PMMar 12
to Albert Kim, Ames Stereo Pipeline Support
This agrees with our own experience. In shadows, SfS cannot give a good result. The sfs_blend solution is a fallback to LOLA, or some prior DEM you have. Gridded LOLA is not that bad at the South Poles recently, as fallback option, if one uses the first dataset mentioned here: https://stereopipeline.readthedocs.io/en/latest/sfs_usage.html#the-initial-terrain

I hope you use many images with diverse illumination, that at least minimizes the regions in darkness.

We usually do not use  --save-dem-with-nodata as this may make blending with LOLA easier than if there is a hole, maybe, not sure. You can try both ways.

(BTW, if you have seams and the result and did not use the more recent recipe in the section "Fixing seams in the SfS terrain" one can try that. I understand this is orthogonal to your issue.)



--
You received this message because you are subscribed to the Google Groups "Ames Stereo Pipeline Support" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ames-stereo-pipeline...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/ames-stereo-pipeline-support/9d277ae1-dd27-40a6-b399-a5c428e7e063n%40googlegroups.com.
Reply all
Reply to author
Forward
0 new messages