CloudCompare forum

Greetings
I have a few questions about Poisson Recon.
1) Starting with a point cloud of about 12 000 000 points, why do when doing the Poisson Recon (octree depth 10) do my mesh finishes with only around 2 000 000 points? Shouldn't the number of points remain the same?

2) Why does the mesh has more faces than vertices? the mesh has around 2 000 000 vertices as mention but around 4 000 000 faces. Shouldn't the faces number be Lower than the point number?

Thank you!

1) The Poisson Surface Reconstruction algorithm creates a surface that is as close as possible to most of the points, while still keeping some smoothness and continuity. It's like covering the points with a drape. Hence, the number of vertices only depends on how detailed the surface is, not on the original points.

2) As each vertex is shared by multiple triangles, it's quite standard for meshes to have more triangles than vertices (even for an algorithm that reuse the original points such as Delaunay triangulation).