Subject: Re: [geometry] 3D box -> 3D multi_polygon conversion
From: Tomislav Maric (tomislav.maric_at_[hidden])
Date: 2013-06-12 04:42:46
On 06/11/2013 11:35 PM, Mateusz Loskot wrote:
> On 11 June 2013 21:12, Adam Wulkiewicz <adam.wulkiewicz_at_[hidden]> wrote:
>> MultiPolygon concept to describe 3D mesh is a good idea. I'd rather provide
>> additional concept.
> I'm no expert of 3D, but isn't there a better concept for 3D mesh storage,
> without redundan vertices and edges, perhaps less topologically fragile than
> multipolygon? (I'm just curious).
You are completely right. The CGAL library uses boost.graph to represent
polyhedra (3D surface mesh) as an edge-graph (half-edge data structure).
This is more general and it allows for optimized operations between
polyhedra: e.g. O(n log n) complexity of intersection between two
polyhedra, where n is the number of points. However, Bruno wrote that
the graph based concept would be much more difficult to implement at
first (if I remember correctly), and the MultiPolygon concept for a
polyhedron is almost already there.
>From my perspective, I am perfectly happy with multipolygon, since the
polyhedra I am aiming at are small (say 10s and not 10000s of faces), so
there is not much data being duplicated. I'm not excluding a graph-based
polyhedron concept, I would just like to work on the multipolygon first,
because I know it works: those 3D algorithms I have up and running for
my Computational Fluid Dynamics (CFD) library.
> Best regards,
Geometry list run by mateusz at loskot.net