[Gmsh] tetrahedralization

[Ares Lagae]

On Wednesday 09 January 2008 03:33:20 pm you wrote:

> Ares Lagae a écrit :
> > I am looking to generate a tetrahedralization of a model.
> > The model is given as a set of triangles, and many of these triangles
> > intersect.
>
> Could you simply tetarhedrize the set of points and then recover some
> triangles ?

The end goal is to use the data in a rendering system, so I cannot afford to 
lose some of the faces.

> Could you repair the triangulation ?
>
> If the model comes from a real cad model (with nurbs & co), then gmsh
> can mesh
> the model in a conforming way (without intersections) and therefore
> generate a tet
> mesh.

The models are triangle soup so I cannot mesh them in a conforming way.

> > Can gmsh generate a volume tetrahedralization of such a model? If I press
> > the "3D" button in the "Mesh" menu, then nothing happens, while other
> > models from the "demo" directory are tetrahedralized.
>
> You should first create a volume with all the faces for gmsh to generate
> the 3D mesh.
> In Gmsh,  any volume should be closed, as it is in TetGen or NetGen. All
> those
> software basically do finite element meshes.

I am not entirely sure what you mean. The input geometry is not necesarily 
manifold or orientable. In many cases there is no clear inside / outside.

The underlying problem is to convert the triangle soup into a simplicial 
complex or piecewise linear complex, which is what e.g. TetGen takes as 
input. The major problem here is eliminating all triangles intersections.

> > If not, is anyone aware of software that can do this? TetGen takes as
> > input a piecewise linear complex, which cannot have intersecting
> > triangles. I had no success with NETGEN.
>
> No, this is pretty much everything I know that is in the public domain.

Thank you for your comments.

-- 
Ares Lagae
Computer Graphics Research Group, Katholieke Universiteit Leuven
http://www.cs.kuleuven.be/~ares/