CGAL users discussion list

["Sebastien Loriot (GeometryFactory)" <>]

Downe, Richard W wrote:
> Ah. I naively assumed that a Nef polyhedron guaranteed no self intersections.
It should, I think in your case this is due to the output precision of 
double.

Sebastien.

> Is there a way to achieve this with Nef polyhedrons? Else...is there a 
> workable technique for pruning self intersections without forfeiting 
> manifoldness?
> -rd
>
> "Sebastien Loriot (GeometryFactory)" 
> <>
>  wrote:
>
>
> Don't know if it's related but you do have a self intersection in the
> .off provided. You can use in the Polyhedron demo the
> "Self intersection " operation to see where it is.
>
> Sebastien.
>
> Richard Downe wrote:
> > I'm using the AABB_polyhedral_oracle surface meshing demo code to remesh
> > a surface -- basically, I combine 2 or more (6 or 7 is the high side of
> > what is common) disjoint tubular meshes using NEF_polyhedron.  The .off
> > file I get when I write the NEF_polyhedron to a file is clean, with no
> > self intersections, and I can view it in tecplot without difficulty.
> >
> > The problem comes with the remeshing -- I remesh to smooth the surface
> > and guarantee that my triangles are as regular as possible, etc., since
> > the downstream processing is computational fluid dynamics.  Most
> > surfaces work fine.  The attached surface I continually get
> >
> > File: /usr/include/CGAL/Surface_mesher/Surface_mesher.h
> > Line: 248
> > Explanation: (0.0200637 0.0244694 0.0292433) is already inserted on
> > surface.
> >
> > in spite of the fact that the .off file appears to be fine.
> > If I set  Tag to Non_manifold, it meshes successfully, but the output
> > surface contains a patch where some of the facets are reversed.  I feel
> > as though this provides some insight into what might be wrong/breaking,
> > but I don't know enough to intelligently pursue it further.  Wondered if
> > anyone else might have insight.
> > Attached is the .off file from the NEF_polyhedron pre-remshing.  I can
> > supply the final mesh as well, but it is quite large, and sending both
> > will put me over the attachments size limit, thus if anyone would like
> > to see it, let me know and I'll send it along separately.
> > Thanks.
> > --rd
>
>
> --
> You are currently subscribed to cgal-discuss.
> To unsubscribe or access the archives, go to
> https://lists-sop.inria.fr/wws/info/cgal-discuss