creating clusters of points based on distance within a bigger set of points

Hi , i have a question. I have a list of points spread out unevenly. Some portions have points that are group closely together and I want to turn them into clusters or new lists based on proximity. In the end there should be several smaller clusters or lists that must form. how do i do that?

Replies to This Discussion

Permalink Reply by Theunis du Toit on October 17, 2009 at 2:21am

Hi Oz,

This is an old post, I remember it from a while ago, have you found a solution yet?

Permalink Reply by Olivier Pasquet on July 5, 2011 at 5:22am

Me too, I'm looking for something that does this. I've looking at Prox but I'm kind of stucked.

Permalink Reply by David Rutten on July 5, 2011 at 5:58am

If I had to do this without scripting I'd create a Delaunay mesh of all the points, then peel the mesh apart into vertices and faces, find the biggest faces and remove those from the face-list. Then you can make a new mesh with the same points but only the small faces and use the Split Disjoint Mesh component to separate them out.

The problem with this approach is that you might 'lose' points. If a point only has large triangles around it, it will disappear from the final set.

I do remember someone implementing an algorithm that finds clusters of points a few months back. The code was posted on this forum (probably in the C#/VB section) but finding it will be tricky.

David Rutten

david@mcneel.com

Poprad, Slovakia

Permalink Reply by Dedackelzucht on July 8, 2011 at 6:14am