Dave,
You are right. The original model is claen and non interesect.
I managed to get to "map" back the triangles based on the method you
mentioned.
Thanks.
jeff
On Jul 24, 1:35 am, "Dave Eberly" <dNOSPAMebe...@[EMAIL PROTECTED]
>
wrote:
> "jeff" <jeff_zhang...@[EMAIL PROTECTED]
> wrote in message
>
>
news:0a921f09-5715-4430-9729-66ed4aa1ffb6@[EMAIL PROTECTED]
>
>
>
> > Hi there,
>
> > I am trying to identify and group a set of triangle based its
> > corresponding geometry model.
>
> > Let's say I have 3 models, a cylinder, a cone and a cube. All these 3
> > models undergone surface me****ng process and generate a common list of
> > the triangles.
>
> > I need to isolate the triangles into a number of group(in this case it
> > would be 3).
>
> > One of the way I am thinking by using the connectivity relation****p
> > i.e start from a vertex of any arbitrary triangle and walking through
> > its neighbors until cover the model.
>
> > Repeat the same process until no more triangles left in the list.
>
> > I wonder if this is the common way to do it or is there any better way
> > to solve this kind of problem?
>
> > There are a lot of unknown related to the connectivity issue. How do I
> > really know that I have finish walk through the whole model.
>
> Are the original models mutually nonintersecting? If so, then you can
> build a collection of edge-triangle meshes, which should themselves be
> mutually nonintersecting. I believe this is what your post hints at.
> Are the triangles stored as triples of vertices? Or do you have triples
> of indices and a common vertex pool? The latter case makes it easier
> to construct the edge-triangle meshes. The former case requires
> identifying equivalent vertices (i.e. you have to build the vertex pool
> from which the vertices are shared).
>
> --
> Dave Eberlyhttp://www.geometrictools.com
|
