algorithm to compute a convex hull
Author
I am trying to figure out given a mesh and the vertex cloud you obtain from it how to build the perfect convex representation from it using only triangles.
Maybe I am just conceptualizing this entirely wrong but anything I envision would end up having triangles overlapping horribly over each other. I really just want my convex hull to be one smooth two-manifold surface. Is there a commonly accepted straightforward algorithm for this?
I am not sure if I have phrased my question right, if anyone's confused I can try to explain using pictures.
Are you familiar with Voronai diagrams, Delauney and Constrained Delauney Triangulation? Delauney Tetrahedralisation? If I understand your description correctly, you may find it useful to study these algorithms and then apply some of the processes involved in 3D. I'm working on a related algorithm for my project.
Are you trying to create convex triangular hulls or convex tetrahedra (which are later reduced to triangles) hulls?
Paulcoz.
Are you trying to create convex triangular hulls or convex tetrahedra (which are later reduced to triangles) hulls?
Paulcoz.
Building convex hulls are _really_ hard. The algorithms themselves are pretty simple, but getting it to work for all point clouds is almost impossible... Even mature libs like qhull can't generate convex hull for some point clouds...
couldn t you start with a bounding box
calculate the distance of each vertex to each plane of the bbox
and use the vertices with the shortest distance to begin with?
calculate the distance of each vertex to each plane of the bbox
and use the vertices with the shortest distance to begin with?
This topic is closed to new replies.
Advertisement
Popular Topics
Advertisement
