Creating a mesh from uneven Floating points

Hi everyone, I am currently trying to figure out a way to create a triangulated mesh from a set of points (which I can change the U and V count) that I got from subsiding a set of oval surfaces,

which I have under the container with the name of Olaf.

After Generating the points from the subdivided uneven surfaces, I organized the point coordinates off of the z value, thinking it would allow me to create links to create a shell like mesh. Problem 1, is that the links between the points are not linking like a triangulated set of lines, and I am getting huge gaps. So, my mesh now has gaps, double surfaces, and disconnected vertices.

Is there a way to organize the coordinates to be able to link the points so that they triangulate to create an exterior shell without having to bake anything?

Attachments:

MODEL 9 21 14.3dm, 1.4 MB
MODEL 9 21 14.gh, 17 KB

Replies to This Discussion

Permalink Reply by David Rutten on September 22, 2014 at 10:39am

You won't be able to triangulate it using a Delaunay approach, as such a triangulation only works in the flat plane and has an outer boundary (the convex hull of the points in the plane). However you can get very close if you can come up with a good mapping from points on the oval to points on the plane. Whether or not it is possible or easy depends on whether there is a single surface underneath all this or perhaps whether the ovals never deviate much from a sphere or ellipsoid.

See this rough sketch of how that might work:

It will work probably best if the distance from origin increases exponentially as points approach the 'gap' or nexus in the oval. Points at the nexus might not even work, as the mapping may specify that they end up at infinity, but that depends on the mathematics. In computing, it's best to stay away from numbers much smaller than 1e-20 and much bigger than 1e+20.

A delaunay mesh topology of the mapped points can be re-applied to the original points, giving you an almost correct result, there'll just be a small hole between the points surrounding the mapping gap.

I don't have time now to delve into specifics (we're entertaining tonight and the food isn't quite done yet), but if this approach sounds like something that might work but you can't work it out yourself, let me know.

Permalink Reply by David Rutten on September 22, 2014 at 1:04pm