CloudCompare forum

Hello,

I have computed geometric features using CloudCompare and also independently with Python. My question revolves around the basis of calculations in CloudCompare. Specifically, I am curious about the formula used for calculating neighbors for Geometric Features, and, more importantly, what criteria are employed to determine the lower limit for the number of neighbors.

I am struggling to find comprehensive documentation on how various features are computed.

Best regards,

Kamilla

Not sure if I understand your question clearly.

The neighbors are extracted via an octree structure. The minimum number of neighbors is generally based on some mathematical considerations (= the minimum information needed to compute the eigenvalues of the covariance matrix for instance).

And the mathematical formulas are better found either in the main articles on which this tool is based: see https://www.cloudcompare.org/doc/wiki/i ... c_features, with the list of articles, especially the one by Hackel et at.