Increasing processor usage?

Sorry for the noob question but… well… I’m a total noob.

I use WebODM to produce orthographic maps of properties my company is selling (real estate company). Typically these are anywhere from 5-50 acres and we end up with 300-600jpg images from the mapping mission.

Running them through the default settings (because I don’t understand any of the settings), these take 4-10 hours it seems and don’t always fully process (as in, it processes a subset of the images but not all of them). I’m running this on a pretty high end machine, but it’s only running at 5-7% CPU usage and like 7gb of ram on a 64gb machine.

Two questions:

  1. How do I increase the CPU and RAM WebODM will use?
  2. How do I manage the datasets that only process part of the set?

A well intentioned Noob.

  1. CPU usage will scale and vary depending upon the part of the pipeline it is in. Not every part is fully multi-threaded. You can ensure it will use up to the maximum number of threads by making sure --max-concurrency is default.
    1A) RAM Usage will vary by image resolution, image count, and settings. If you’re running default settings, I would not expect to see all 64GB being used ever (unless you’re processing an absolutely massive dataset).
  2. Partial reconstruction is almost always due to sub-optimal flight planning and data collection. You can try to increase --min-num-features, disable Image Resizing, increase --feature-quality, and potentially disable pre-matching of images.
1 Like

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