RE: [Algorithms] merging of points in 3D space

[Richard F. <alg...@th...>]

1) if you are having a problem find out whether they are close enough to
merge, may I suggest you use a dynamic octree for finding out if they
are close, then adding an entry into the octree if it is a new vert...

a dynamic octree is very fast at picking out points in space...

2) if you are having problems deciding where to put the merged vertex,
then why not put it on the grid point where the distance between grid
points is the tolerance of the merge.

using the grid you can put a definite stop on the leaf size of your
octree...

if you use a grid, there is no reason why you cannot use a hash table
(might be quicker than an octree)

------------------------------------------------------------------------
-------
main(){char*a="main(){char*a=%c%s%c;printf(a,34,a,34);}";printf(a,34,a,3
4);}
------------------------------------------------------------------------
-------


> -----Original Message-----
> From: gda...@li... 
> [mailto:gda...@li...] On 
> Behalf Of Peter Szinek
> Sent: 18 April 2003 10:26
> To: gda...@li...
> Subject: [Algorithms] merging of points in 3D space
> 
> 
> Hello all,
> 
> I am having the following problem:
> 
> Consider an array of points in 3D given by (x, y, z). Every 
> point is in the
> array more times (5-6 times, does not really matter) but the 
> problem is that
> these points are not really the same - there are small 
> differences between their
> coordinates
> e.g. points with coordinates
> 
> 1.000 1.000 1.000
> 0.999 1.001 1.000
> 1.000 1.002 0.989 etc
> 
> are considered to represent the same point.
> 
> Now, my question is, how to 'merge' these points into one 
> (e.g. in the previous
> example, keep only one such point) _VERY_ fast. The trivial 
> algorithm is far-far
> away (as usual) from the fast enough.
> 
> Any help appreciated.
> 
> 
> 
>   
> 
> -- 
> Best regards,
>  Peter                          mailto:sz...@po...
> 
> 
> 
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf
> _______________________________________________
> GDAlgorithms-list mailing list
> GDA...@li...
> https://lists.sourceforge.net/lists/listinfo/gdalgorithms-list
> Archives:
> http://sourceforge.net/mailarchive/forum.php?forum_id=6188
>