Hi, I am building a program which has to read in a set of 3d coordinates from a file and render these in 3d space. This is fine and I have done it with no problem, however as these points are on the surface of a solid object I would like to turn them into polygons and render that instead. Has anyone heard of any algorithms which do this? Thanks Ian

It is called (3D) surface reconstruction and there are lots of algorithms for doing it (none particularly simple though). I don''t have any handy links but google should give some good hits.

For convex surfaces, Delaunay triangulation and/or Voronoi tessellation can be used. I can''t really understand what you''re trying to do, so I''ll just point out the general rule:

• if you want to minimze polygon size (and, inversely, maximize polygon count), run the mesh through a Delaunay triangulator
  
• if you don''t care about polygon complexity (the number of corners they have), run it through a Voronoi tessellator
  

Both are somewhat challenging to implement and will require you to know your way around math. However, both allow you to control the number of polygons you want to split the mesh into by choosing the number and location of a random set of reference points (called "sites"), so you can control the process and its results pretty well.

---