Page 1 of 1

Radius "r" for calculating cloud density

Posted: Mon Nov 16, 2015 10:47 pm
by asghar
I was wondering how the default (suggested) 'r' radius for cloud density is calculated. I tried to find out how it is calculated based on each point cloud characteristics but could not figure out. Any help/ clue will be much appreciated!

Re: Radius "r" for calculating cloud density

Posted: Tue Nov 17, 2015 8:16 am
by daniel
For the density we use a very coarse estimation of the cloud surface (based on the bounding-box volume!). The we pick the radius that would theoretically give 12 points per square units. See the code of MainWindow::GetDefaultCloudKernelSize in 'mainwindow.cpp'.

This is only a very first guess, very approximate. The user should know better which radius is best!

Re: Radius "r" for calculating cloud density

Posted: Wed Nov 18, 2015 11:16 pm
by asghar
Thanks for your reply, Daniel. That was very helpful.

One more question, is that the same radius that the program uses for generating sample points from a CAD file, say .stl, (using Edit>>Mesh>>Sample Points). Also , it looks like there is a density limit of 1,000,000 for the generated point cloud. Can this limit be increased to create a denser point cloud?

Re: Radius "r" for calculating cloud density

Posted: Thu Nov 19, 2015 9:21 am
by daniel
The 'Mesh > Sample points' feature expects either a number of point (and there's no real limit for this number apart for your computer memory) or a density.

The density is not a radius (it's a number of points per square units). The density is indeed limited to 1 M. points per square units but there's no real reason for this (we will increase it). Meanwhile you can use the other method (absolute number of points). You can compute the mesh surface first (with 'Edit > Mesh > Measure surface') if you need a precise specific density.