ODM_Filterpoints seems to always be a huge bottleneck when running big datasets. Hopefully in the future someone can figure this out. Unfortunately I’m unable to contribute with coding on this.
Only one core seems to be in use when on a multi core system, causing big costs when running big datasets on the cloud.
Not sure if this is “OK” here, but I just found out that WebODM or ODM is on BountySource.com
I have added $50 dollars towards parallelizing ODM_Filterpoints at some point.
Thank you.
Hey @korn704 
thanks for posting a bounty! Everything helps.
Time ago I started a branch to parallelize filterpoints, but never finished it.
https://github.com/pierotofy/OpenDroneMap/commit/4b02a47728a5ade1ab95026aaab85395fae99e27
If somebody picks it up it would be awesome. I don’t have a timeline for completing it myself, but it’s certainly a bottleneck we need to improve.
Bump!
https://fund.webodm.org/#/fund/1
Everyone, please consider helping with this. This is something that would help all us shortening the processing time on large data sets. Running on the cloud would help tremendously with costs. Definitely a bottleneck when processing large datasets with systems with very high amount of cores (64,96, etc)
Thank you!
Yes, I think this is incredibly important as a first step towards improving the processing speed of WebODM and have joined in.
Added my support. Thanks for moving this forward.
Threw a little in myself, as my latest dataset was >50,000,000 points.