Number of images not necessarily related to memory usage or time taken

Only 25 photos taken while holding the drone in my hand

25 images 04:57:27
|Options:|auto-boundary: true, debug: true, feature-quality: ultra, gps-accuracy: 6, mesh-octree-depth: 13, mesh-size: 600000, min-num-features: 15000, optimize-disk-space: true, orthophoto-resolution: 0.1, pc-geometric: true, pc-quality: ultra, resize-to: -1, skip-orthophoto: true, use-3dmesh: true, verbose: true|
|Average GSD:|0.06 cm|
|Area:|7.06 m²|
|Reconstructed Points:|35,941,003|

Normally I’d expect 25 images on ultra/ultra to complete in less than 20 minutes, but here around 100% of 96GB RAM was in use for over 2 1/2 hours, with the swap disk working very hard during that time too!

Processing started at about 5h 5m.

I’ve done much larger datasets on similar settings in much less time, what could cause the behaviour seen here?

1 Like

My hunch: GSD estimates and super weirdness associated therein resulting in weird memory nonsense (waves hand in vague way).

4 Likes

In other words: figure out what it is setting for a GSD from the logs. Then set the GSD so 10 or 100x that value in a rerun. It should complete. Un-referenced datasets can result in poorly parameterized models with impossible memory requirements. Usually not, just occasionally.

1 Like

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.