CGAL users discussion list

["Tomislav Maric" <>]

Well, the mesh is hexahedral and arbitrary. To be sure, by arbitrary I mean 
this:

- mesh is a set of cells with boundary patches (lists of labels (IDs) to the 
total face list, corresponding to the boundary faces)
- a cell is a list of labels referring to an array of faces
- a face is a list of (counterclockwise counted) labels referring to an array 
of points

The mesh connectivity is computed accross the faces: there is an 
owner-neighbour relationship which is defined by the cell ID (needed for the 
communication: which cell needs to communicate data to which Polyedron_3 
point, and vice-versa). The cells with lower IDs "own" the faces. Each face 
represents a boundary between max two cells.

This kind of construction is used as an unstructured mesh used for a Finite 
Volume numerical method solution. In its basis, the mesh is consisted of 
waterproof (logical: faces are not checked for planarity) polyhedra, but I am 
interested only in hexagonal cell geometry, for now.

The boxes are axis alligned, because the cells of the mesh are axis alligned.

Because I need to perform communications between the Polyhedron_3 and this 
Mesh, I'm thinking in the direction of an implementation of the 
BoxIntersectionBox_d concept, that will inherit from the "cell" class of the 
Mesh and provide the requirements needed by the concept. This way, I can 
re-use the info from the Mesh, and call CGAL algorithms on it. Either this, 
or build the Box from the array of points of each cell (min, max) + cell ID. 


Thanks a lot! 

Tomislav


> ----- Original Message -----
> From: Sebastien Loriot (GeometryFactory)
> Sent: 02/28/12 02:45 PM
> To: 
> 
> Subject: Re: [cgal-discuss] Polyhedron_3 points / Array of boxes
> 
> How is you hexa-mesh? is it regular or arbitrary? are the boxes 
> axis-aligned?
> 
> A screenshot can be helpful too if you have one.
> 
> Sebastien.
> 
> On 02/28/2012 11:22 AM, Tomislav Maric wrote:
> > Hi everyone,
> >
> > what would be the fastest way to compute something like "point in box" 
> > for the points of the Polyhedron_3? I have an array of boxes (extremely 
> > large, well over hundreds of thousands) and a rather large Polyhedron_3 
> > depicting a surface mesh. What I am aiming at is a fast way of 
> > communicating data between an underlying hexahedral mesh (array of boxes) 
> > and the vertices of the Polyhedron..
> >
> > I've read the chapter on Intersecting sequences of d-Dimensional boxes, 
> > and I'm thinking about this:
> >
> > 1) Create an array of boxes for the Polyhedron_3 facets (triangular mesh)
> > 2) Create an array of boxes for the hex mesh.
> > 3) Compute the intersection test.
> > 4) Sub process: go through every triangle, take the point of the triangle 
> > and find the box (out of the candidates given by the intersection test) 
> > within which the point is located.
> >
> > This sounds a bit... slow... so any advice is appreciated! :)
> >
> > Thanks!
> >
> > Tomislav
> >
> 
> 
> -- 
> You are currently subscribed to cgal-discuss.
> To unsubscribe or access the archives, go to
> https://lists-sop.inria.fr/wws/info/cgal-discuss