CGAL users discussion list

[Sebastien Loriot <>]

It is still not clear to me what you are looking for.
If you have 3D points, getting all the facets (triangles) will give you
all the facets on the convex hull + some facets internal to the convex hull.
Are you more looking into something like point cloud -> triangle mesh?

In such a case have a look here:
https://doc.cgal.org/latest/Manual/tuto_reconstruction.html

Best,

Sebastien.

On 7/5/22 18:30, KL ( via cgal-discuss Mailing List) 
wrote:
> Well,
> I am not even sure about the wording anymore, too complicated.
> I don't even understand which API I should use.
> The context is 3D triangulation from an array of 3D points.
> A high level triangulation function -> inject your points and get the 
> minimum number of triangles from those, would be very handy.
> This is the only thing I want.
> Iterating over vertices is not good either because you don't have the 
> logic of the triangles.
>  From there, I may create a 3D object.
> Hence the associated constraint to have an optimized number of resulting 
> triangles.
> So, Sebastian, where should I look to achieve that ?
> I am a lost sailor,
> --
> MKL
> 05.07.2022, 17:21, "Sebastien Loriot ( via 
> cgal-discuss Mailing List)" <>:
>
>     I'm not sure to understand what is your problem. A minimal self
>     contained example that we can run and that shows your problem would
>     be perfect.
>
>     What I can tell from the code I see:
>     You are computing a 3 triangulations that consists of tetrahedra.
>     `finite_facets()` provides the range of face that are not connected
>     to the infinite vertex. A facet is given a std::pair of a Cell_handle
>     (tetrahedron) and the id of the vertex in the cell opposite to the
>     facet. About the access to the vertices of a facet, the %4 is here so
>     that the index stays in [0,3] while ignoring pair.second (that might or
>     might not be infinite).
>
>     HTH,
>
>     Sebastien.
>
>     On 7/5/22 14:46, KL (
>     <> via cgal-discuss Mailing List)
>     wrote:
>
>           Hi,
>           The issue: while the code below works, it returns too many
>         triangles for
>           a simple shape.
>           So, for a square, instead of two triangles made by one of the
>         diagonal
>           of the square, I get four tris, because the two diagonals are
>         connecting.
>           I get my vertices using:
>           using K = CGAL::Exact_predicates_inexact_constructions_kernel;
>           using Triangulation = CGAL::Triangulation_3 <K>;
>           using Cell_handle = Triangulation::Cell_handle;
>           using Vertex_handle = Triangulation::Vertex_handle;
>           using Locate_type = Triangulation::Locate_type;
>           using Point_CGAL = Triangulation::Point;
>           ...
>           Triangulation tri;
>           ...
>           // load data
>           ...
>           //loop
>           for ( auto &e : tri.finite_facets() ){
>                    auto &vert = *( e.first );
>                    auto starter_id = e.second;
>                    for ( u32 x = 1; x < 4; x += 1 ){
>                            auto &pt = *( vert.vertex( ( starter_id + x )
>         % 4 )
>           );                 // saw that line on the stackoverflow -> 3
>         finite
>           vertices and 1 infinite vertice, hence the 4.
>                            // store my result
>                    }
>           }
>           I need the triangulation to be the minimum number of tris for
>         any given
>           shape. A square should have 2 tris.
>           I read the doc, but it is not among the easiest to wrap one's
>         head around.
>           What should I modify, please?
>           Thanks
>           --
>           MKL
>
>           --
>           You are currently subscribed to cgal-discuss.
>           To unsubscribe or access the archives, go to
>         https://sympa.inria.fr/sympa/info/cgal-discuss
>         <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
>     <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