CGAL users discussion list

["Sebastien Loriot (GeometryFactory)" <>]

Orient polygon soup will duplicate vertices when some non-manifold 
situations are encountered.
polygon_soup_to_polygon_mesh() creates a mesh out of a correctly 
oriented soup representing a manifold mesh.

The merging of "duplicated" vertices can be done using stitching.
Right now we do nothing for non-manifold situtations but we have
some pull-request that will improve it.

Sebastien.

On 08/16/2018 11:06 PM, 

 wrote:
> I see, thanks for the clarification. After merging the duplicated 
> vertices the non-manifold edges become evident. Incidentally, I though 
> the input polygon soup could be represented redundantly by storing an 
> instance of a vertex in each incident triangle, and that merging the 
> vertices would be done by polygon_soup_to_polygon_mesh(). So from my 
> understanding, this function "only" takes care of building the data 
> structure and of explicitly setting the adjacency relations between its 
> elements - correct?
>
> Claudio
>
> On Thu, Aug 16, 2018 at 12:12 PM Sebastien Loriot (GeometryFactory) 
> < <mailto:>> wrote:
>
>     As is your model is detected as valid surface mesh.
>     This is because you have many duplicated vertices.
>     If you merge them (using something like [1] for example)
>     you will end up with a mesh with only one boundary cycle.
>     It is not stitch-able because it is would produce a set of non-manifold
>     edges.
>
>       From there, depending on what you want to do with the model, you can
>     still try to close the model but you'll endup with some
>     self-intersection.
>
>     Sebastien.
>
>     [1] https://gist.github.com/sloriot/12f6ba2c5d593b1b0e2e82a07fc82df7
>
>     On 08/15/2018 11:44 PM, 
>
>     
> <mailto:>
>  wrote:
>      > Hi Sebastien,
>      >
>      > thank you for looking into this. I've exported the polygon soup into
>      > .off (see attachment) - let me know if you need any further
>     information.
>      >
>      > Claudio
>      >
>      > On Wed, Aug 15, 2018 at 12:29 PM Sebastien Loriot (GeometryFactory)
>      > 
> <
>  
> <mailto:>
>     
> <mailto:
>  
> <mailto:>>>
>  wrote:
>      >
>      >     You probably have some non-manifold faces in your soup. If
>     this is the
>      >     case, it probably needs some clean up. Can you share the
>     input so that
>      >     we can give more info?
>      >
>      >
>      >     Sebastien.
>      >
>      >
>      >
>      >     On 08/14/2018 10:05 AM, 
>
>     
> <mailto:>
>      >     
> <mailto:
>  
> <mailto:>>
>  wrote:
>      >      > Hello,
>      >      >
>      >      > I'm trying to use the CGAL's PMP package to create a mesh
>     out of a
>      >      > polygon soup (extracted from a spatial subdivision
>     structure using a
>      >      > min-cut approach). I've usedorient_polygon_soup() and
>      >      > polygon_soup_to_polygon_mesh(). A subsequent call to
>      >      > is_polygon_soup_a_polygon_mesh() return true, but the mesh
>      >     produced is
>      >      > not closed. This also creates problems when using other
>     functions
>      >     of the
>      >      > package, such as
>      >      >
>      >      > is_outward_oriented()/reverse_face_orientations(). Also using
>      >      > stitch_borders did not improve the mesh.
>      >      >
>      >      > Is this the expected behaviour of the functions?
>      >      >
>      >      > Thanks,
>      >      >
>      >      > Claudio
>      >      >
>      >      >
>      >      >
>      >
>      >     --
>      >     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