Hints about any slim method, numerical or not, to quickly identify areas of pointclouds with different density as illustrated below.
In the picture there are two pointclouds, the upper half and the lower half, where the lower half has lower point density (not very visible, but anyway).
I would like to be able to quickly identify the two different areas (inside the circle) based on a density constraint and finally find the centroid of the respective areas that would emerge. Lets assume that the point distribution in each area are somewhat consistent, meaning that in this case there are not more than two different “densities” in the cloud.
In the picture the red circle would represent what I’m looking for, the area-center of the densest area (which in this case is the upper half).
If this was in image processing, I’d expand and blur the dots until they blended into smooth regions of different shades of grey, then threshold this.
Interesting approaches! Thank you for the hints. I’ll examine them carefully.
Although working on a very specific problem just now I will later probably go in the direction of working with image analysis, or convert some problems to images, and continue some of the problem solving from there. For now I have my hands dirty with pointclouds as the data source.
But I first calculate a surface associated to each point using Voronoi, each surface is converted to a density (1/m²) then this density is affected to the point, …
