Efficient 3D box cropping

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

Efficient 3D box cropping

EWER115
This post was updated on .
Hi all!

I have a very large point clouds with 30M+ points and need to slice them to
a regular structures (voxelization) let's say to 5x5x5m boxes. I have
precomputed each slice min_pt and max_pt and used getPointsInBox for each of
the slice. This process is extremely slow as for each box the method
iterates through all the points in the cloud. Is there any efficient
implementation of this procedure?

Best, Dejan



--
Sent from: http://www.pcl-users.org/
_______________________________________________
PCL-users@pointclouds.org / http://pointclouds.org
http://pointclouds.org/mailman/listinfo/pcl-users
Reply | Threaded
Open this post in threaded view
|

Re: Efficient 3D box cropping

Richard Frank
You could look into dividing the overall space into a grid ( octree perhaps)
And consider the approach as finding the box for the point rather than the point(s) for the box. Then you’re visiting each point once and depending on the data structure, a relative small search for the box ( voxel)


Rick Frank
Dominion Software, Inc.
825 Beacon Street
Newton, MA 02459
Medical, Scientific, and Industrial Software

> On Jan 31, 2018, at 6:03 AM, EWER115 <[hidden email]> wrote:
>
> Hi all!
>
> I have a very large point clouds with 30M+ points and need to slice them to
> a regular structure (voxelization) let's say to 5x5x5m boxes max. I have
> precomputed each slice min_pt and max_pt and used getPointsInBox for each of
> the slice. This process is extremely slow as for each box the method
> iterates through all the points in the cloud. Is there any efficient
> implementation of this proccedure?
>
> Best, Dejan
>
>
>
> --
> Sent from: http://www.pcl-users.org/
> _______________________________________________
> [hidden email] / http://pointclouds.org
> http://pointclouds.org/mailman/listinfo/pcl-users
_______________________________________________
[hidden email] / http://pointclouds.org
http://pointclouds.org/mailman/listinfo/pcl-users