CGAL users discussion list

[Ben McWhorter <>]

Hi Sebastian,

The reason that I test if they are disjoint before joining is because  
I need each disjoint nef polyhedron to be output separately.

Also, the problem that I was having is that when I would join them  
together without the check, I would get resulting nef polyhedron that  
were either not simple or not valid.

Doing the join the way I do, gives me outputs that are all simple and  
valid.  I don't believe that you will have a problem with those 3  
examples I sent in particular, but simply joining all of my input  
together in the manner you did gives me a result that is either not  
simple or not valid.

Does that make more sense?

If you want me to show you an example of the join producing this  
invalid output, I can find some example input that displays the problem.


Thanks,
Ben

On Jun 15, 2011, at 12:51 PM, Sebastien Loriot (GeometryFactory) wrote:

> Sorry but still not sure to get your problem.
>
> I try to make the union of output3.off, output4.off and output5.off
> in one run, with a (what I call minimal) example joint. I don't have
> time to understand all your code but why are you testing if nef are
> disjoint before making the union?
> Joining disjoint nef polyhedra is handled.
>
> Anyway, I may be wrong but to me it seems that the need of the two  
> runs
> come from the way you handle the union.
>
>
> Sebastien.
>
> #include <CGAL/Exact_predicates_exact_constructions_kernel.h>
> #include <CGAL/Polyhedron_3.h>
> #include <CGAL/IO/Polyhedron_iostream.h>
> #include <CGAL/Nef_polyhedron_3.h>
> #include <CGAL/IO/Nef_polyhedron_iostream_3.h>
> #include <CGAL/OFF_to_nef_3.h>
> #include <iostream>
> #include <fstream>
>
> typedef CGAL::Exact_predicates_exact_constructions_kernel  Kernel;
> typedef CGAL::Polyhedron_3<Kernel>  Polyhedron;
> typedef CGAL::Nef_polyhedron_3<Kernel> Nef_polyhedron;
>
> int main(int, char *argv[]) {
>  std::ifstream off1(argv[1]);
>  std::ifstream off2(argv[2]);
>  std::ifstream off3(argv[3]);
>  Nef_polyhedron n1,n2,n3;
>  CGAL::OFF_to_nef_3(off1,n1);
>  CGAL::OFF_to_nef_3(off2,n2);
>  CGAL::OFF_to_nef_3(off3,n3);
>
>  n1 += n2;
>  n1+=n3;
>
>  Polyhedron p;
>  n1.convert_to_polyhedron(p);
>  std::cout << p << std::endl;
> }
>
>
> Benjamin McWhorter wrote:
> > Hi Sebastian,   Thanks again for your help.  The following .cpp  
> > file is what I have been using.  I am reattaching the files I sent  
> > in the last email with it.  The print_polyhedron_wavefront isn't  
> > really too important, it takes only a polyhedron input (so I do use  
> > convert_to_polyhedron), it is just convenient to get the output in  
> > a .obj file so I can look at it in meshlab.  This program takes a  
> > directory name as input, where the .obj files lie.  It creates .OFF  
> > files for each .obj and then tries to join them.  The first time  
> > you run with the attached 3 .obj files, you will get 2 .obj files  
> > as output.  If you then run it again with those two output .obj  
> > files, you will get one output.
> > What I was hoping is that I would get one output from the first  
> > time I ran the program.
> > Thanks again,
> > Ben
> > ------------------------------------------------------------------------
> > On Jun 15, 2011, at 11:38 AM, Sebastien Loriot (GeometryFactory)  
> > wrote:
> > > Not sure to understand what you want.
> > >
> > > 1)You want to make 1+2+3 but you can't because it seems that the  
> > > union
> > > is not possible when object are disjoint. if so could you provide  
> > > 1,2
> > > and 3 as off (a private message only for the files is better or a  
> > > URL).
> > >
> > > 2)I do not know the function print_polyhedron_wavefront (which is  
> > > not
> > > documented), I am using the member function convert_to_polyhedron.
> > > However it may happen that some degenerate faces appear when  
> > > converting
> > > exact coordinates to double. Do you have the same result when using
> > > convert_to_polyhedron?
> > >
> > > Providing a minimal example is very appreciated, I then just need to
> > > compile it without having to create it from scratch (and this  
> > > ensures
> > > we have the same setting).
> > >
> > > Sebastien.
> > >
> > > Benjamin McWhorter wrote:
> > > > Hi Sebastian,
> > > > Sorry if I wasn't clear on my problem.  I have attached my  
> > > > original input .obj files and the resulting converted .off files.
> > > > The problem that I am having is this, while I get an intersection  
> > > > and a resulting join that is valid between 2 of the 3, CGAL does  
> > > > not find a valid and simple join between the result of that join  
> > > > and the other nef_polyhedron (hopefully thats clear).  My process  
> > > > for performing the join is like this:
> > > > convert .obj to .off
> > > > create nef_polyhedron from .off using the off_to_nef3 constructor.
> > > > test if there is an intersection between the two nef_polyhedron
> > > > if there is test if the result of a join between the two would  
> > > > give a simple and valid result
> > > > if so, join them.
> > > > I then convert back to polyhedron and write the output to a file  
> > > > using print_polyhedron_wavefront
> > > > The problem is that running the program on the attached 3 inputs  
> > > > gives 2 resulting nef_polyhedron.  If I run the program again,  
> > > > (using the output from the first run as input), then it will join  
> > > > the two polyhedron (after discarding the facets you mentioned  
> > > > when constructing the Nef_polyhedron).
> > > > So I basically have two questions -- why is it that these  
> > > > nef_polyhedron do not get joined in the first place?
> > > > Second -- why is CGAL's output (from print_polyhedron_wavefront)  
> > > > is giving me Polyhedron with degeneracies, when I do not get that  
> > > > problem from the attached input -- only CGAL produced input.
> > > > I hope that all makes sense and as always I very much appreciate  
> > > > your help.
> > > > Thanks,
> > > > Ben
> > > > ------------------------------------------------------------------------
> > > > On Jun 15, 2011, at 3:36 AM, Sebastien Loriot (GeometryFactory)  
> > > > wrote:
> > > > > If you are using the construction from polyhedron,
> > > > > nef assumes that the input is a valid polyhedron.
> > > > > Your input output4.OFF contains facets made only of collinear
> > > > > points.
> > > > >
> > > > > The following code detect these facets:
> > > > >
> > > > > for (Polyhedron::Facet_iterator f=poly.facets_begin();
> > > > >   f!=poly.facets_end();++f)
> > > > > {
> > > > >  Kernel::Point_3 pt1,pt2,pt3;
> > > > >  pt1=f->halfedge()->vertex()->point();
> > > > >  pt2=f->halfedge()->next()->vertex()->point();
> > > > >  Polyhedron::Halfedge_handle h=f->halfedge()->next();
> > > > >  do{
> > > > >    pt3=f->halfedge()->next()->next()->vertex()->point();
> > > > >    if (! CGAL::collinear(pt1,pt2,pt3) ) break;
> > > > >    h=h->next();
> > > > >  }
> > > > >  while(h!=f->halfedge());
> > > > >  if( h==f->halfedge() )
> > > > >    std::cerr<<"Error degenerate facet" << std::endl;
> > > > > }
> > > > >
> > > > > In order to remove this kind of degeneracies, you can use the
> > > > > function CGAL::OFF_to_nef_3 documented here:
> > > > > http://www.cgal.org/Manual/latest/doc_html/cgal_manual/Nef_3_ref/Function_OFF_to_nef_3.html
> > > > >
> > > > > Using this function I was able to make the union of your
> > > > > inputs with CGAL-3.8.
> > > > >
> > > > > Sebastien.
> > > > >
> > > > >
> > > > > Benjamin McWhorter wrote:
> > > > > > Hello All,
> > > > > > I have been having some trouble correctly joining  
> > > > > > Nef_Polyhedron.  I have attached some .off files that should  
> > > > > > recreate the error.  I have been trying to join a large number  
> > > > > > of Polyhedron using CGAL and ran into many problems along the  
> > > > > > way.  I found that I cannot simply join the simple and valid  
> > > > > > nef_polyhedron and expect a simple and valid result.  The only  
> > > > > > way I was able to accurately join them was by testing if the  
> > > > > > result of the join was simple and valid after checking if there  
> > > > > > was an intersection and if it passed all three tests, then  
> > > > > > perform the join.
> > > > > > It seems that the order in which I joined them had a lot to do  
> > > > > > with it working out.
> > > > > > Anyhow, on to my actual issue.  I ran the join and expected the  
> > > > > > following two polyhedron to be joined, but they were not.   
> > > > > > After the join I convert back to Polyhedron and then use the  
> > > > > > print_polyhedron_wavefront function to write the output in  
> > > > > > an .obj file.
> > > > > > These .OFF files that I attached are simply  
> > > > > > print_polyhedron_wavefront .objs that I have converted  
> > > > > > to .OFF.  These two were not joined during execution, however,  
> > > > > > when I ran the program again with these two as input, I get the  
> > > > > > following (sorry if I put too much).  I am wondering why CGAL  
> > > > > > is discarding facets that were created by the  
> > > > > > print_polyhedron_wavefront function and 2, why these two were  
> > > > > > not joined in the first place.  Any help is much appreciated.
> > > > > > Thanks,
> > > > > > Ben
> > > > > > ------------------------------------------------------------------------
> > > > > > output4.OFF
> > > > > > Conversion from vertex cycle to Nef_polyhedron_3 was not  
> > > > > > successful. Error history: -> function parameter 'normal' is  
> > > > > > NULL_VECTOR (this can be a symptom of an error). Now, direction  
> > > > > > of projection is 'x'. -> Same vertex appears multiple in cycle.  
> > > > > > Now, direction of projection is 'y'. -> Same vertex appears  
> > > > > > multiple in cycle. Now, direction of projection is 'z'. -> Same  
> > > > > > vertex appears multiple in cycle. Finally, empty  
> > > > > > Nef_polyhedron_3 was constructed.
> > > > > > Hence, discard input facet 14 (enumerated beginning with 1).  
> > > > > > Check semantics!
> > > > > > Conversion from vertex cycle to Nef_polyhedron_3 was not  
> > > > > > successful. Error history: -> function parameter 'normal' is  
> > > > > > NULL_VECTOR (this can be a symptom of an error). Now, direction  
> > > > > > of projection is 'x'. -> Same vertex appears multiple in cycle.  
> > > > > > Now, direction of projection is 'y'. -> Same vertex appears  
> > > > > > multiple in cycle. Now, direction of projection is 'z'. -> Same  
> > > > > > vertex appears multiple in cycle. Finally, empty  
> > > > > > Nef_polyhedron_3 was constructed.
> > > > > > Hence, discard input facet 21 (enumerated beginning with 1).  
> > > > > > Check semantics!
> > > > > > Conversion from vertex cycle to Nef_polyhedron_3 was not  
> > > > > > successful. Error history: -> function parameter 'normal' is  
> > > > > > NULL_VECTOR (this can be a symptom of an error). Now, direction  
> > > > > > of projection is 'x'. Now, direction of projection is 'y'. Now,  
> > > > > > direction of projection is 'z'. -> Different vertices are  
> > > > > > projected to same location! Finally, empty Nef_polyhedron_3 was  
> > > > > > constructed.
> > > > > > Hence, discard input facet 29 (enumerated beginning with 1).  
> > > > > > Check semantics!
> > > > > > Conversion from vertex cycle to Nef_polyhedron_3 was not  
> > > > > > successful. Error history: -> function parameter 'normal' is  
> > > > > > NULL_VECTOR (this can be a symptom of an error). Now, direction  
> > > > > > of projection is 'x'. -> Same vertex appears multiple in cycle.  
> > > > > > Now, direction of projection is 'y'. -> Same vertex appears  
> > > > > > multiple in cycle. Now, direction of projection is 'z'. ->  
> > > > > > Different vertices are projected to same location! Finally,  
> > > > > > empty Nef_polyhedron_3 was constructed.
> > > > > > Hence, discard input facet 30 (enumerated beginning with 1).  
> > > > > > Check semantics!
> > > > > > Conversion from vertex cycle to Nef_polyhedron_3 was not  
> > > > > > successful. Error history: -> function parameter 'normal' is  
> > > > > > NULL_VECTOR (this can be a symptom of an error). Now, direction  
> > > > > > of projection is 'x'. -> Same vertex appears multiple in cycle.  
> > > > > > Now, direction of projection is 'y'. -> Same vertex appears  
> > > > > > multiple in cycle. Now, direction of projection is 'z'. ->  
> > > > > > Different vertices are projected to same location! Finally,  
> > > > > > empty Nef_polyhedron_3 was constructed.
> > > > > > Hence, discard input facet 51 (enumerated beginning with 1).  
> > > > > > Check semantics!
> > > > > > output4.OFF is simple
> > > > > > output4.OFF is valid
> > > > > > output3.OFF
> > > > > > output3.OFF is simple
> > > > > > output3.OFF is valid
> > > > > --
> > > > > 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
>
>
> --
> You are currently subscribed to cgal-discuss.
> To unsubscribe or access the archives, go to
> https://lists-sop.inria.fr/wws/info/cgal-discuss