CGAL users discussion list

Okay, thanks for your help.
Here is my minimalistic example:

#include <CGAL/Simple_cartesian.h>
#include <CGAL/Polyhedron_3.h>

typedef CGAL::Simple_cartesian<double> Kernel;
typedef CGAL::Polyhedron_3<Kernel> Polyhedron;
typedef Polyhedron::Halfedge_handle Halfedge_handle;
typedef Kernel::Point_3 Point_3;
typedef Polyhedron::Facet_iterator Facet_iterator;
typedef Polyhedron::Halfedge_iterator Halfedge_iterator;

Polyhedron polyhedron;
ifstream fin  ("model.off");
fin >> polyhedron;

for ( Halfedge_iterator j = polyhedron.halfedges_begin(); j != 
polyhedron.halfedges_end(); )
{
   Halfedge_iterator i = j;
   j++;
   if( i->is_border_edge() )
      continue;         
   Polyhedron::Facet_handle f1 = i->facet();
   Polyhedron::Facet_handle f2 = i->opposite()->facet();
   Halfedge_handle he1 = f1->halfedge();
   
   Point_3 p0 = he1->vertex()->point();
   Point_3 p1 = he1->next()->vertex()->point();
   Point_3 p2 = he1->next()->next()->vertex()->point();

   Kernel::Vector_3 n_i = CGAL::cross_product(p0-p2, p1-p2);
   n_i = n_i /sqrt(n_i.squared_length());
   
   Halfedge_handle he2 = f2->halfedge();
   p0 = he2->vertex()->point();
   p1 = he2->next()->vertex()->point();
   p2 = he2->next()->next()->vertex()->point();

   Kernel::Vector_3 n_j = CGAL::cross_product(p0-p2, p1-p2);
   n_j = n_j /sqrt(n_j.squared_length());

   if(n_i == n_j)
   {
      if( circulator_size(i->opposite()->vertex_begin()) >= 3 && 
circulator_size(i->vertex_begin()) >= 3 )
      {
         ++j;
         polyhedron.join_facet(i);
      }
      else
         std::cout << "faces could not be joined\n";
   }
}

The model is consiting of 4 triangles in a plane, so I would expect a quad as 
output, instead I get 2 remaining triangles.

Here is the model in .off format:

OFF
5 4 0
-100 -100 0
100 -100 0
100 100 0
-100 100 0
0 0 0
3 3 0 4
3 4 2 3
3 1 4 0
3 2 4 1

Greetings, 
Joel



Andreas Fabri schrieb:
> Joel,
>
> it would be helpful, if you saved the polyhedron just before the join 
> operation fails.
> Ideally, you provide a minimalistic example, and you tell us on which edge, 
> that is
> which two points, it fails.  Then we can try to have a look.
>
> andreas
>
> 
>  wrote:
>> Thanks for the help, the iteration works now.
>> Sorry, I dont really know what an antenna is in the context of a 
>> polyhedron.
>>
>> I will try to clarify what I want to do:
>> My input mesh is a triangle mesh, for example a cube.
>> This cube consists of 12 triangles, 2 for each side.
>> What I want to have is a polyhedron with 6 polygons (quads in this case).
>>
>> For the simple example of the cube it works well, but when I want to do 
>> the same with a triangulated quad consisting of many triangles, there are 
>> many edges which can not be joined via polyhedron.join_facet(...).
>> I dont really know the reason for that, it would be nice if someone could 
>> explain that.
>>
>> Joel
>>
>>
>> -------- Original-Nachricht --------
>>> Datum: Wed, 06 May 2009 21:48:52 +0200
>>> Von: "Dr. J.K. Becker" 
>>> <>
>>> An: 
>>> 
>>> Betreff: Re: [cgal-discuss] Joining faces in polyhedron
>>
>>> I am not sure if I really understood this Mathieu, but would that not be
>>> an antenna which CGAL really does not like at all anyway (CGAL is happy
>>> to produce one, but any further operations on a polyhedron with an
>>> antenna will fail, at least it does for me)? And no, as far as I know,
>>> there are no build in checks for antennas...
>>> Anyway, polyhedra (or polygons) can have holes, no problem with that in
>>> CGAL.
>>> Joel, I am not sure I understood what you want to do correctly, but if
>>> you want to join all incident facets of a 3D triangulation that have
>>> the same normal, you will not join anything really (at least you will
>>> most likely never end up with a closed polyhedron)?! Except if you
>>> started of with a regular grid of points for the triangulation, but in
>>> that case you dont really need to triangulate anything? I am trying to
>>> figure out something very similar (and have not solved it) so maybe I am
>>> completely wrong....
>>>
>>> Jens
>>>  
>>>
>>> On Wed, 2009-05-06 at 18:08 +0200, Mathieu Brédif wrote:
>>>> Hi,
>>>>
>>>> If I am not mistaken, polyhedral facets may not have holes in CGAL.
>>>> This means that you can only remove edges if they are not adjacent to
>>>> the same face on both sides.
>>>>
>>>> Mathieu
>>>>
>>>> On Wed, May 6, 2009 at 5:52 PM, Andreas Fabri
>>>> <>
>>>>  wrote:
>>>>> Hi Joel,
>>>>>
>>>>> You must make a copy of the iterator, increment it and then work on
>>> the copy
>>>>> in the body of the loop.
>>>>> This is because the halfedges are in an in-place list, your join
>>> operation
>>>>> removes them and the iterator
>>>>> no longer points on something reasonable
>>>>>
>>>>> for ( Halfedge_iterator j = polyhedron.halfedges_begin(); j !=
>>>>> polyhedron.halfedges_end(); ){
>>>>>  Halfedge_iterator i = j;
>>>>>  ++j;
>>>>>  // and now your code again
>>>>>  .
>>>>>  .
>>>>> }
>>>>>
>>>>>
>>>>> I think you have even to increment j twice, as the twin halfedges are
>>> stored
>>>>> consecutively.
>>>>> Alternatively you should use an Edge_iterator.
>>>>>
>>>>>
>>>>> andreas
>>>>>
>>>>>
>>>>> 
>>>>>  wrote:
>>>>>> Hi!
>>>>>>
>>>>>> I am trying to convert a triangle mesh to a polyhedron, i.e. I want
>>> to
>>>>>> join all faces which are adjacent and have the same normal
>>>>>>
>>>>>> As an easy example I chose a flat patch consisting of triangles in
>>> .off
>>>>>> format. As a result I want to have a polyhedron consisting of a
>>> single
>>>>>> polygon.
>>>>>>
>>>>>> I was able to load the file, but the follwoing did not work:
>>>>>>
>>>>>> for ( Halfedge_iterator i = polyhedron.halfedges_begin(); i !=
>>>>>> polyhedron.halfedges_end(); ++i) {
>>>>>>    if( i->is_border_edge() )
>>>>>>        continue;
>>>>>>
>>>>>>    Polyhedron::Facet_handle f1 = i->facet();
>>>>>>    Polyhedron::Facet_handle f2 = i->opposite()->facet();
>>>>>>        Halfedge_handle he1 = f1->halfedge();
>>>>>>        Point_3 p0 = he1->vertex()->point();
>>>>>>    Point_3 p1 = he1->next()->vertex()->point();
>>>>>>    Point_3 p2 = he1->next()->next()->vertex()->point();
>>>>>>        Kernel::Vector_3 n_i = CGAL::cross_product(p0-p2, p1-p2);
>>>>>>    n_i = n_i / sqrt(n_i.squared_length());
>>>>>>
>>>>>>    Halfedge_handle he2 = f2->halfedge();
>>>>>>    p0 = he2->vertex()->point();
>>>>>>    p1 = he2->next()->vertex()->point();
>>>>>>    p2 = he2->next()->next()->vertex()->point();
>>>>>>
>>>>>>    Kernel::Vector_3 n_j = CGAL::cross_product(p0-p2, p1-p2);
>>>>>>    n_j = n_j / sqrt(n_j.squared_length());
>>>>>>
>>>>>>    if(n_i == n_j)
>>>>>>    {
>>>>>>        if( circulator_size(i->opposite()->vertex_begin()) >= 3 &&
>>>>>> circulator_size(i->vertex_begin()) >= 3 )
>>>>>>            polyhedron.join_facet(i);
>>>>>>    }
>>>>>> }
>>>>>>
>>>>>> There seem to be 2 problems: - after joining to faces the iterator
>>> becomes
>>>>>> invalid
>>>>>> - some faces cannot be joined, but why and how can I change that?
>>>>>>
>>>>>> Can anyone help me please?
>>>>>>
>>>>>> Have a nice day,
>>>>>> Joel
>>>>> -- 
>>>>> You are currently subscribed to cgal-discuss.
>>>>> To unsubscribe or access the archives, go to
>>>>> https://lists-sop.inria.fr/wws/info/cgal-discuss
>>>>>
>>> -- 
>>> You are currently subscribed to cgal-discuss.
>>> To unsubscribe or access the archives, go to
>>> https://lists-sop.inria.fr/wws/info/cgal-discuss
>>
>
>
>
>
> Eingehende eMail ist virenfrei.
> Von AVG überprüft - www.avg.de 
> Version: 8.5.325 / Virendatenbank: 270.12.19/2099 - Ausgabedatum: 05/05/09 
> 13:07:00 
>

-- 
Psssst! Schon vom neuen GMX MultiMessenger gehört? Der kann`s mit allen: 
http://www.gmx.net/de/go/multimessenger01