Poisson Reconstruction Skip the Triangulation stage + Does MLS take note of my original normals?

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view
|

Poisson Reconstruction Skip the Triangulation stage + Does MLS take note of my original normals?

DevSH Graphics Programming on Gmail
Hi,

I have two questions, let me start with the easier.

"Does MLS take note of my original normals?"

Essentially I have to use MeshLab and not PCL (unfortunately) as I have
multiple scans or flyovers without a single sensor position.
MeshLab has that cool graph based feature to orient normals consistently.
If I wanted to use MLS to smooth my data, would it f**up my normals?
I know it fits a polynomial surface to the points and gets normals from
that, but how does it know which side of the surface is the "outside"?
Will it take my original normal as seed?


Second Question:
"How do I get the reconstructed isofunction from the Poisson filter?"


I want to dump the isofunction to a uniform voxel grid, and I don't
want/don't care about the iso-surface.

Also it seems I must know the computed iso-value at which the surface
would have been.

I know that the Poisson surface reconstruction filter has this function
evaluated over a pcl::poisson:Octree<> as it does marching cubes over it.

How could I query this octree at arbitrary points for the values of the
iso-function?
How is the iso-function interpolated across the cells/nodes?
_______________________________________________
[hidden email] / http://pointclouds.org
http://pointclouds.org/mailman/listinfo/pcl-users