CGAL users discussion list

[Benjamin Kehlet <>]

2014-10-14 9:55 GMT+02:00 Jane Tournois 
<>:
> Hi Benjamin,
>
> Le 13/10/2014 20:01, Benjamin Kehlet a écrit :
>
> Hi Jane
>
> 2014-10-13 17:25 GMT+02:00 Jane Tournois
> <>:
>
> Hi Benjamin,
>
> you are not doing anything wrong.
>
> As you have noticed, the mesh itself is the C3T3
> (Complex_3_in_triangulation_3).
> The C3T3 is a subset of the triangulation.
> So, there can be cells, and vertices, that belong to the triangulation, but
> not to the C3T3.
>
> In your particular example, the criteria that you are using to generate the
> mesh are a bit too weak. One symptom of that is that the surface is not very
> well fitted by the mesh. You can see it if you output the mesh to Medit for
> example.
>
> About isolated vertices, it's the same as for cells. It is not forbidden to
> have one or several vertices that belong to the triangulation but not to the
> triangulation.
> It can be the case for example when a vertex is first inserted on the
> boundary (let's say during the initialization step), but then the refinement
> process changes the connectivity around it, and the vertex can start not to
> belong to the boundary anymore.
>
> Ah! I didn't know that could happen.
>
> To make it short :
> - it is not a bug
> - the object that are interested in at the end/during the mesh generation
> process is the C3T3
>
> Thanks a lot for the explanation, Jane! That was very enlightening.
>
> So just to be sure: It is then my responsibility to grab only the
> relevant vertices from the triangulation, right? (Since the C3T3 does
> not provide an iterator over the vertices that belong to the complex.)
>
> Yes, you are right.
>
> You can iterate on cells, using the Cells_in_complex_iterator
> (see
> http://doc.cgal.org/latest/Mesh_3/classMeshComplex__3InTriangulation__3.html
> )
> and then collect the vertices.
>
> The iterator on vertices that you can use on a C3T3 is
> Vertices_in_complex_iterator
> (see
> http://doc.cgal.org/latest/Mesh_3/classMeshComplexWithFeatures__3InTriangulation__3.html
> )
> but it iterates only on the corners of the object.

Excellent! Thanks a lot for your help!

Benjamin

>
> Regards,
> Jane.
>
>
>
> Regards
>
> Benjamin
>
> Best regards,
>
> Jane.
>
>
> --
> Jane Tournois, PhD
> R&D Engineer at GeometryFactory
> http://www.geometryfactory.com/
>
>
>
> Le 13/10/2014 10:41, Benjamin Kehlet a écrit :
>
> 2014-10-13 9:55 GMT+02:00 Clément Jamin 
> <>:
>
> Hi Benjamin,
>
> Do you use the *parallel* mesh algorithm that is available in CGAL 4.5
> (with CGAL::Parallel_tag)?
>
> Hello Clément!
>
> No, this is in serial.
>
> Regards
>
> Benjamin
>
> Best,
> Clément.
>
>
> On Mon, Oct 13, 2014 at 12:44 AM, Benjamin Kehlet 
> <>
> wrote:
>
> Hello CGAL community!
>
> When meshing polyhedral domains, I sometimes experience isolated
> vertices, ie. vertices which does not belong to any cell. The attached
> file demonstrates the problem. Could someone take a look, and check if
> I'm doing anything wrong? The code for reading out the mesh is based
> on the code in CGAL for exporting to medit's file format.
>
> I get this result when attempting to mesh the file
> demo/Polyhedron/data/rotor.off from CGAL's source tree:
>
> benjamik@benjamik-ThinkPad:~/tmp/cgal-meshing/build$ ./main
> ~/software/CGAL-4.5/demo/Polyhedron/data/rotor.off
>
> [...]
>
> Number of unconnected vertices:14
> 0.0693619 0.340288 -0.260398 0
> 0.148692 0.344385 -0.252795 0
> 0.11059 0.332426 -0.267569 0
> 0.0800736 0.308173 -0.290017 0
> 0.0358875 0.312982 -0.283919 0
> 0.0739249 0.2902 -0.306196 0
> 0.100469 0.304801 -0.293518 0
> 0.0949104 0.318775 -0.276573 0
> 0.0777684 0.350724 -0.248457 0
> 0.0909964 0.338994 -0.261008 0
> 0.0263695 0.298269 -0.299189 0
> 0.00269295 0.307881 -0.290388 0
> 0.0496229 0.300013 -0.297675 0
> -0.0248875 0.311644 -0.285615 0
>
> Also a few other files (at least cow.off and bones.off and maybe
> others) in the same directory reproduces the problem.
>
> Best regards
>
> Benjamin Kehlet
>
> --
> You are currently subscribed to cgal-discuss.
> To unsubscribe or access the archives, go to
> https://sympa.inria.fr/sympa/info/cgal-discuss
>
>
> --
> You are currently subscribed to cgal-discuss.
> To unsubscribe or access the archives, go to
> https://sympa.inria.fr/sympa/info/cgal-discuss
>
>
>
>
> --
> You are currently subscribed to cgal-discuss.
> To unsubscribe or access the archives, go to
> https://sympa.inria.fr/sympa/info/cgal-discuss
>
>
>
>
> --
> Jane Tournois, PhD
> R&D Engineer at GeometryFactory
> http://www.geometryfactory.com/