Subject: CGAL users discussion list
List archive
- From: blandc <>
- To:
- Subject: [cgal-discuss] Convex decomposition of a cube (should be a cube)
- Date: Thu, 3 Sep 2015 07:45:41 -0700 (PDT)
- Authentication-results: mail3-smtp-sop.national.inria.fr; spf=None ; spf=None ; spf=None
- Ironport-phdr: 9a23:biSLxRCehkQvWybyDUMcUyQJP3N1i/DPJgcQr6AfoPdwSP7+psbcNUDSrc9gkEXOFd2CrakU16yO7+u/CCQp2tWojjMrSNR0TRgLiMEbzUQLIfWuLgnFFsPsdDEwB89YVVVorDmROElRH9viNRWJ+iXhpQAbFhi3DwdpPOO9QteU1JTrkbjusMKCKyxzxxODIppKZC2sqgvQssREyaBDEY0WjiXzn31TZu5NznlpL1/A1zz158O34YIxu38I46FpwecbD/S/JvxwDecAVHV1e1wyscbkvB2GQQqU7WYHSU0XlABJCk7L9kLURJD05yH+v+xx02HOONXxSLkuXT2kx7ZiSBjvziwAcSM6pjKEwvdshb5W9Uvy7yd0xJTZNdrNbPc=
Greetings.
I'm trying to do convex decompositions of an arbitrary triangulated mesh
(arbitrary but with nice properties like closed orientable manifold) using
CGAL's convex_decomposition_3. This means I read the mesh, convert to a
Nef_polyhedron, and call the decomposition function. I'm essentially using
the list_of_convex_parts.cpp code, except for how I get the nef polyhedron.
This works, but tends to produce (I think) unnecessarily large number of
components if the mesh has edges adjacent to faces that form an (exact) 180
degree angle. This is the case for something as simple as a cube where each
cube face is split in two triangles by a diagonal edge. Rather than return a
single component (a cube is convex after all), I get 5 components plus the
exterior.
(I also tried to obtain convex components of meshes I know are convex but
have no faces forming an 180 angle, like Blender's icosphere; I obtain only
one convex component in such cases, which is as expected.)
Is there some preprocessing I can/should do to a Nef polyhedron (obtained
from an arbitrary triangulated mesh) to avoid this? Thanks!
Cheers!
--
View this message in context:
http://cgal-discuss.949826.n4.nabble.com/Convex-decomposition-of-a-cube-should-be-a-cube-tp4661126.html
Sent from the cgal-discuss mailing list archive at Nabble.com.
- [cgal-discuss] Convex decomposition of a cube (should be a cube), blandc, 09/03/2015
Archive powered by MHonArc 2.6.18.