I can the opensfm features directory after feature extraction on the docker drive. I can’t find the matches and tracks. Does opensfm on odm use a different dataset structure than listed at Dataset Structure — OpenSfM 0.4.0 documentation?
The feature files appear in the feature directory in real time as odm is running, but I don’t see any changes in the project directory while the matches are running.
After 3 days of running the matching part of opensfm the matches directory appeared. I was expecting it to be like feature extraction where a file is written for each image processed.
Since my previous attempts had always failed halfway through the directory was never made.
Since FLANN matching is the longest step I may open am issue with OpenSfM to see if they cam did partial writes and restart processing halfway through. This could also be useful if just a few new images were added.