search

mkazhdan / PoissonRecon

Poisson Surface Reconstruction
MIT License
1.59k stars 431 forks source link

Consult about SurfaceTrimer #215

Open SalvatoreDamon111 opened 2 years ago

SalvatoreDamon111 commented 2 years ago

Hi,Professor ! I'm recently looking at Poisson surface reconstruction, and I'd like to know more about the implementation steps of SurfaceTrimmer as you mentioned, I hope you can explain in as much details as possible.Looking forward to your reply！THANKS

mkazhdan commented 2 years ago

Can you be more concrete about what it is you are trying to understand? E.g. "How is it implemented?", "What do particular parameters mean?", etc.

SalvatoreDamon111 commented 2 years ago

hi professor , Actually, I saw the operation related to SurfaceTrimmer on https://github.com/mkazhdan/PoissonRecon, we did not discuss it before, I just simply want to know how is it implemented step by step, Just describe it briefly,, what are the steps . thank you very much, Demon Have a nice life!

mkazhdan @.***> 于2022年3月13日周日 08:36写道：

Can you be more concrete about what it is you are trying to understand? E.g. "How is it implemented?", "What do particular parameters mean?", etc.

— Reply to this email directly, view it on GitHub https://github.com/mkazhdan/PoissonRecon/issues/215#issuecomment-1065992727, or unsubscribe https://github.com/notifications/unsubscribe-auth/AYGBBQL2NROPAGV2SZILKMLU7U2BZANCNFSM5QRVSGTA . Triage notifications on the go with GitHub Mobile for iOS https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675 or Android https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub.

You are receiving this because you authored the thread.Message ID: @.***>

mkazhdan commented 2 years ago

The idea is reasonably straight-forward.

As part of the reconstruction process PR needs to estimate the sampling density at every point. Assuming that points live within the unit cube, this essentially amounts to assigning a density-depth value, d, to each point, where 1/2^d is the average spacing between the point and its neighbors. (Roughly, this is used to determine how deep into the octree the sample should be placed. E.g. if the sampling density is too coarse, the octree may not be refined to the depth prescribed by the "--depth" parameter.)

When running PR you can have the executable output the estimated depth at each mesh vertex using the "--density" flag.

Then, you can run the surface trimmer executable to remove the parts of the reconstructed surface where the estimated density-depth was smaller than some threshold.