It is possible to generate a mesh from a cloud of nodes?

User Help
#1

Basically that, instead of using triangle or tetgen directly use MMG. Maybe not with the current API, but calling some code directly?

Thanks in advance

Can mmg3d be used to turn an unconstrained to a constrained Delaunay tetrahedralization?
#2

Hi, I don’t see any other answers here yet, but I don’t think this is possible with mmg.

You have to generate a base surface mesh first (somehow) from the cloud of points. I use MeshLab, it works pretty well.

Then generate a base tet mesh using Netgen or something.
Then use MMG to alter and improve the tet mesh.

That’s the only way I know of. Others may have different or better solutions.
Best, T.

#3

Hi @loumalouomega , hi @tcdoe,

Actually it is not possible with Mmg:

  • first, the delaunay kernel of Mmg insert points only in the middle of mesh edges (so it needs to be modified to insert points everywhere);
  • second, as explained by @tcdoe, the enforcement of the surface mesh of the cloup point doesn’t exist in Mmg.

The proposed solution seems pretty good. Otherwise, if the searched surface mesh is the convex hull of the point clouds, I thik that you can directly use the tetgen software to generate the surface and volume mesh.

Best regards,
Algiane

#4

Hi,

The recent post of @Mathias-Fuchs (Can mmg3d be used to turn an unconstrained to a constrained Delaunay tetrahedralization?) make me think to a way to genereate a volume mesh from a point cloud with Mmg: If you are able to define and compute a signed distance function on your point cloud, you can use the -ls option of Mmg to discretize explicitely the 0-value of the distance function and to obtain a mesh of the point cloud.

Best Regards,

Algiane