CGAL users discussion list

[Simon Giraudot <>]

Hello,

There is no notion of planar/crease/corner points in advancing front 
(the reference paper is "A greedy delaunay-based surface reconstruction 
algorithm" by Cohen-Steiner & Da). You may be confusing with the point 
set structuring algorithm that is used as a pre-processing in this 
example: 
https://doc.cgal.org/latest/Advancing_front_surface_reconstruction/index.html#title8

If you want to get inspiration from this example for your own dataset, 
you could imagine creating a Priority functor that would give a higher 
priority to crease faces (I'm not sure how this fits your data).

Note that the Scale Space algorithm could help handling complicated data 
set by applying advancing front on a smooth version of your dataset. 
Again I'm not sure how it applies to your data, but it might be worth 
the trial:
https://doc.cgal.org/latest/Scale_space_reconstruction_3/Scale_space_reconstruction_3_2scale_space_advancing_front_8cpp-example.html

Best,

--
Simon Giraudot, PhD
R&D Engineer
GeometryFactory - http://geometryfactory.com/

Le 14/03/2021 à 22:07, HDYNE ( via 
cgal-discuss Mailing List) a écrit :
> Hi all,
>
> I am developing a dedicated cad application for sheet metal processing; I am
> taking a DXF and discretizing it (lines, arcs and circles) to a point cloud.
> The aim is to build a 3D model where several operations can be performed
> including folding (ideally using the Triangulated Surface Mesh Deformation
> https://doc.cgal.org/latest/Surface_mesh_deformation/index.html).
>
> I am currently using the advancing front surface reconstruction algorithm to
> build my 3D mesh
> (https://doc.cgal.org/latest/Advancing_front_surface_reconstruction/index.html),
> however this fails (the mesh is far from correct) as soon as the part that I
> am meshing is more complicated than just a square. I have read the paper
> discussing this algorithm and they introduce the idea of planar, crease and
> corner points. I was wondering how do I go about specifying a point to be a
> crease? Since the DXF is a cut file, every point generated by the
> discretization will be a crease and I am hoping that specifying this will
> generate much more ideal results. The model doesn't have to be perfect as
> this is simply a representation, the output file will be a modified DXF.
>
> I would also like the output to work with the triangulated mesh deformation
> as this is the next step in this application.
>
> Any help would be much appreciated, or any suggestions for another way that
> I can build a 3D model from a DXF would be of great help also.
>
> I have attached some test files that have discretized:
>
> hard_dxf.xyz
> <http://cgal-discuss.949826.n4.nabble.com/file/t376291/hard_dxf.xyz>
> medium_dxf.xyz
> <http://cgal-discuss.949826.n4.nabble.com/file/t376291/medium_dxf.xyz>
> easy_dxf.xyz
> <http://cgal-discuss.949826.n4.nabble.com/file/t376291/easy_dxf.xyz>
>
> Many regards,
>
> Cheers, Hayden.
>
>
>
> --
> Sent from: http://cgal-discuss.949826.n4.nabble.com/