Description
An algorithm which reconstructs a water-tight triangulated surface, minimizing a criterion of that surface, just from point coordinates, has been presented in this poster and documented in this thesis.Meshes, which are generated by modeling, or reconstructed from sensor devices, often exhibit topological defects, e.g. triangle flaps with boundary edges, intersecting triangles and so on (see also this Eurographics 2012 tutorial). For simulation, further modeling or searching it is often required that they fulfill certain properties, such as being manifold or even water-tight.
Since the above method constructs a surface with the desired properties from a starting set which is not manifold either, completing that algorithm on meshes with any kind of defects should guarantee equivalent results for that minimization objective, without having to resort to handling specific cases.
Tasks
- Fit the mesh into a constrained Delaunay triangulation, e.g. by subdividing or bounding non-Delaunay triangles.
- Adapt the above method to start with an existing triangle set.
- Extend it to handle huge meshes out-of-core.
- Compare results with existing mesh repairing methods
