CGAL users discussion list

[Michael Jeulin-L <>]

Laurent Saboret wrote:
> Michael Jeulin-L wrote:
> > Laurent Saboret wrote:
> > > Michael Jeulin-L wrote:
> > > > Laurent Saboret wrote:
> > > > > Michael Jeulin-L wrote:
> > > > > > Laurent Saboret wrote:
> > > > > > > Hi Michael,
> > > > > > >
> > > > > > > The assertion below might occur when the Poisson reconstruction 
> > > > > > > parameters are out of the recommended range.
> > > > > > > See "Contouring parameters" paragraph at 
> > > > > > > http://www.cgal.org/Manual/3.5/doc_html/cgal_manual/Surface_reconstruction_points_3/Chapter_main.html#Subsection_45.6.1 
> > > > > > >
> > > > > > >
> > > > > > > You may reconstruct your objects with various parameters with 
> > > > > > > the examples provided with CGAL:
> > > > > > > - poisson_reconstruction in 
> > > > > > > examples/Surface_reconstruction_points_3 (command line)
> > > > > > > - Point_set_demo in demo/Surface_reconstruction_points_3 (Qt GUI)
> > > > > > > I recommend you to try first with the default parameters, then 
> > > > > > > with the parameters you used in your code.
> > > > > > >
> > > > > > > Best regards,
> > > > > > > Laurent Saboret
> > > > > > >
> > > > > > >
> > > > > > > Michael Jeulin-L wrote:
> > > > > > > > Hello everybody,
> > > > > > > >
> > > > > > > > Excuse me to disturb, but I have a problem with Poisson 
> > > > > > > > function and I cannot
> > > > > > > > find the solution to fix the problem.
> > > > > > > >
> > > > > > > > From this example :
> > > > > > > > http://www.cgal.org/Manual/3.5/doc_html/cgal_manual/Surface_reconstruction_points_3/Chapter_main.html 
> > > > > > > >
> > > > > > > >
> > > > > > > > I try to compute several object with a loop, for some of them 
> > > > > > > > all is working
> > > > > > > > fine and I have the output I expect. However, for others my 
> > > > > > > > application crash
> > > > > > > > with the following error :
> > > > > > > >
> > > > > > > > CGAL error: precondition violation!
> > > > > > > > Expression : tester(d)
> > > > > > > > File : CGAL/Triangulation_3.h
> > > > > > > > Line: 643
> > > > > > > > Explanation:
> > > > > > > > Refer to the bug-reporting instructions at 
> > > > > > > > http://www.cgal.org/bug_report.html
> > > > > > > >
> > > > > > > > when I try to compute, I think : 
> > > > > > > > "function.compute_implicit_function()"
> > > > > > > > When I look in the file triangularization_3.h, it seems to be 
> > > > > > > > caused by a
> > > > > > > > conflict with cell. How could I resolved the problem before my 
> > > > > > > > application bug
> > > > > > > > ?
> > > > > > > >
> > > > > > > > Thank you very much for every answer.
> > > > > > >   
> > > > > > Thank you very much.
> > > > > >
> > > > > > But I mean, I try to compute the surface for very different kind 
> > > > > > of objects.
> > > > > > So, is there a way to know before computing until the crash if 
> > > > > > the conditions are matched or not.
> > > > > > It`s the same problem for the normals computations.
> > > > > >
> > > > > > No way to make a preprocessing in order to detect wrong objects ?
> > > > > >
> > > > > > Actually I don`t think so, but still ask the question just in case.
> > > > > >
> > > > > > Thank.
> > > > > > Kind Regards.
> > > > > Hi Michael,
> > > > >
> > > > > The examples provided with CGAL (see above and also 
> > > > > examples/Point_set_processing_3 for normals computation) compute 
> > > > > appropriate parameters for any input object.
> > > > >
> > > > > Best regards,
> > > > > Laurent Saboret
> > > > Hi,
> > > >
> > > > Indeed, It is the case for normals computation but not for 
> > > > polyhedrons one.
> > > > So, I still have the problem to prevent/avoid the assertions when 
> > > > the Poisson reconstruction parameters are out of the recommended 
> > > > range.
> > > >
> > > > Is there any way to do that or should I change directly CGAL source 
> > > > code ?
> > > >
> > > > Thank you very much.
> > > > Best Regards.
> > > Hi Michael,
> > >
> > > I am not sure that I understand your question. The examples above 
> > > compute appropriate parameters for the reconstruction function 
> > > make_surface_mesh(). This should be enough to reconstruct your data.
> > > Am I missing something?
> > >
> > > Best regards,
> > > Laurent Saboret
> > Hi,
> > The problem is that`s not. Cause when I compute the parameters I have 
> > in some cases assertion error.
> > It could be to compute average_spacing, bounding_sphere... and if I 
> > have understand well Poisson Construction (
> >
> > // Defines the implicit surface: requires defining a
> >      // conservative bounding sphere centered at inner point.
> >
> > ), I need this parameters to compute the implicit surface.
> >
> > So, in function of my options :
> >
> > // Poisson options
> >    FT sm_angle = 20.0; // Min triangle angle in degrees.
> >    FT sm_radius = 30; // Max triangle size w.r.t. point set average 
> > spacing.
> >    FT sm_distance = 0.375; // Surface Approximation error w.r.t. 
> > point set average spacing.
> >
> > +
> >
> >  FT average_spacing = CGAL::compute_average_spacing (points.begin(), 
> > points.end(),
> >                                                       6 /* knn = 1 
> > ring */);
> >
> > My autommatic reconstruction works better or worse but still bug at a 
> > moment.
> > I just try to be able to prevents this bug cause by out of range 
> > parameters for Poisson Reconstruction.
> >
> >
> > Thank you to interest to my problem.
> > Kind Regards.
> Hi Michael,
>
> Please post the code and an input file that triggers your problem.
>
> Best regards,
> Laurent Saboret
>
>
> Ce courrier électronique et toutes les pièces éventuellement jointes 
> qu’il contient sont CONFIDENTIELS et destinés exclusivement à l’usage 
> de leur destinataire. Si une erreur de transmission ou une adresse 
> erronée a mal dirigée ce courrier, merci d’en informer l’expéditeur en 
> lui faisant une réponse par courrier électronique dès réception. Si 
> vous n’êtes pas le destinataire de ce courrier, vous ne devez pas 
> l’utiliser, le conserver, en faire état, le distribuer, le copier, 
> l’imprimer ou en révéler le contenu à une tierce partie.
> Ce courrier électronique est à usage strictement informatif et ne 
> saurait engager de quelque manière que ce soit INFOTERRA France SAS, 
> ni ses filiales.
>
> This e-mail and any attachments hereto are CONFIDENTIAL and intended 
> solely for the use of the addressee. If you have received this e-mail 
> in error please send it back to the person that sent it to you.
> If you have received it in error, please notify the sender by return 
> email. If you are not the addressee of this email, you must not use, 
> keep, disseminate, copy, print or otherwise deal with it.
> This email is for information only and will not bind INFOTERRA France 
> SAS in any contract or obligation, nor its subsidiaries.
My cause which cause problem is more or less exactly the same that we 
can find as exemple on the last manual documentation of CGAL :

/void PBuilding::ComputeImplSurfaces(double Angle, double Radius, double 
Distance, int NbNeighbors)
{
   int index=0;

   _completion=0                 ;
   int size=_pointsNormals.size();
   global_bar->setMaximum(size)  ;

   // Poisson options
   double angle    = Angle   , // Min triangle angle in degrees.
          radius   = Radius  , // Max triangle size w.r.t. point set 
average spacing.
          distance = Distance, // Surface Approximation error w.r.t. 
point set average spacing.
          average_spacing ,    // Average Spacing between Points
          sphere_radius   ,
          dichotomy_error ;    // Must be << than distance

   Point       inner_point;
   Sphere     bSphere;
   Polyhedron Building;

   for(PointVecList::iterator house=_pointsNormals.begin(); 
house!=_pointsNormals.end(); ++house)
   { 
       global_bar->setValue(_completion++);

       STr tr;            // 3D Delaunay triangulation for surface 
generation
       C2t3 c2t3(tr);  // 2D Complexe in 3D Delaunay trinangulation
          
       Poisson_Reconstruction_Function function(
                             house->begin(), house->end(),
                             
CGAL::First_of_pair_property_map<PointVectorPair>(),
                             
CGAL::Second_of_pair_property_map<PointVectorPair>());

       if (!function.compute_implicit_function()) continue;

       average_spacing = CGAL::compute_average_spacing(house->begin(), 
house->end(), CGAL::First_of_pair_property_map<PointVectorPair>(), 
NbNeighbors);
       inner_point        = function.get_inner_point();
       bSphere            = function.bounding_sphere();
       radius            = sqrt(bSphere.squared_radius());
       sphere_radius    = 5.0*radius;
       dichotomy_error = distance*average_spacing/1000.0;

       Surface_3 surface(function,
                       Sphere(inner_point, sphere_radius*sphere_radius),
                       dichotomy_error/sphere_radius);

       CGAL::Surface_mesh_default_criteria_3<STr> criteria(angle, 
radius*average_spacing, distance*average_spacing);
       CGAL::make_surface_mesh(c2t3,                                  
// reconstructed mesh
                               surface,                              // 
implicit surface   
                               criteria,                              
// meshing criteria   
                               CGAL::Manifold_with_boundary_tag());  // 
require manifold mesh
      
       CGAL::output_surface_facets_to_polyhedron(c2t3, Building);

       string file = "PoissonRecon\\Bat" + 
boost::lexical_cast<std::string>(index++) + ".off";
           ofstream out(file.c_str());
           out << Building;

       _polyhedronsDl.push_back(Building);
       Building.clear();
   }

   global_bar->setValue(size);
   global_bar->reset();
}

/And the problem can come for different part of the parameters computation.
For instance the last one was :

CGAL error: precondition violation!
Expression : tester(d)
File : CGAL/Triangulation_3.h
Line: 643
Explanation:
Refer to the bug-reporting instructions at 
http://www.cgal.org/bug_report.html

But it could an other assertion just like I was saying.
Do you know how can I prevent that (density calculation, number of 
points.... to be able to say : This Normalized Points will be out of 
range for Poisson Reconstruction).

Thanks.
Kind Regards.

--
--------------------------------------------
Michael Jeulin-L, ComputaMaps

Private Bag X20, Constantia 7848, South Africa
tel: +27.21.700 8140, fax: +27.21.701 9520, cell: +27.72.8588181
e-mail: 


Manufacturers of digital maps and imagery
http://www.computamaps.com
--------------------------------------------