[vtkusers] Problem with surface reconstruction from cloud point, vtkDelaunay3D works?

[Kevin H. Hobbs]

On Wed, 2005-10-26 at 21:38 +0800, Mr.Y wrote:
>    Thanks for your help.
>    But, to my knowledge vtkDelaunay3D uses some  algrythm like convex hull  to construct a surface, and its 
> basic function is to construct a triangular nets from unstructed points, but My problem is first to pick out the 
> points lies on the surface from a points cloud, which has points scattered inside it. And then to construct a both 
> convex and concave surface from the picked points.
>    Anyway, I would try vtkDelaunay3D if it do works.
>     Is there some more suggestions?


If the length of edges allowed is low, then the result will differ from
"the convex hull." I think this sort of consideration exists for most
algorithms.

These points,
 
     o o
    
     o o
  
 
o o  

o o

could be triangulated this way

     o-o
    /|\|
   / o-o
  / / /
 / / /
o-o / 
|\|/
o-o

but if edges are only allowed to be one character long then even the
topology could be different.

     o-o
     |\|
     o-o

 
o-o 
|\|
o-o
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part
URL: <http://www.vtk.org/pipermail/vtkusers/attachments/20051026/77896db5/attachment.pgp>