Re: [Algorithms] Heightfield to NURBS conversion

[Christian R. M. K. <chr...@pu...>]

Martin Fuller wrote:
> 
> What you're asking for is a easy solution to what is a very very complex
> problem. I would suggest instead of using NURBS looking at a curve 
> equation that actually interpolates the control points. (but break the 
> convex hull) Catmull Rom splines immediatly spring to mind.

Unfortunately it MUST be NURBS. The project I participate in uses them 
and there is no way around that. And it MUST be approximating since I need 
fewer control points than there are samples in the heightfield.

Christian