Hi everybody,
I need to recreate a similar behaviour than what the Pop3d (Populate 3d) component does in Grasshopper, but in RhinoPython for a project.
I’m basically looking for a way to populate a box with random, three-dimensional points.
Now, I have no need for perfectly distributed, random sampling, but the formation of too many distinct point clusters should be prevented.
I’ve implemented Poisson Disc Sampling/blue noise sampling before, but in two dimensions? Does anybody know if it scales well into 3D?
Another idea was to subdivide the box into a three-dimensional grid of sub-regions, corresponding to the total number of sample points. Each 3D cell would contain a single point that could be jittered for randomness. I guess a neat by-product here would be the tree-like data structure that would make nearest neighbour searches fast and easy.
What do you think? Have you come a cross something more straightforward?
Any suggestion?
8 posts - 5 participants