The geometrical model and the ranscac modules are very usefull to fit different object to a point cloud even in very noisy case.
However the geometrical shapes that implemented up to now are only two: Palane and Sphere.
I have implemented an ellipsoid model and its ransac version.
The fitting algorithm is the least suqares, both to fit all the points in the cloud and a subset of them.
The minimum number of non complanar points to perform the fit is 11.
Description
The geometrical model and the ranscac modules are very usefull to fit different object to a point cloud even in very noisy case. However the geometrical shapes that implemented up to now are only two: Palane and Sphere. I have implemented an ellipsoid model and its ransac version.
The fitting algorithm is the least suqares, both to fit all the points in the cloud and a subset of them. The minimum number of non complanar points to perform the fit is 11.
References
least squares fitting for the ellipsoid : https://github.com/aleksandrbazhin/ellipsoid_fit_python