A Novel Method of Cone Fitting Based on 3D Point Cloud

摘要

The accuracy and stability of circular conical surface is easily influenced by the different distribution of noise and points in point cloud, when 3D point cloud data fits conical surface. For this reason, this paper proposes a novel method of establishing and optimizing the objective function with conical vertex. First, calculate the axis direction and half-apex angle of the cone; second, calculate conical vertex according to the result got in the first step; third, establish objective function with axis direction and half-apex angle and vertex, and optimize it with Levenberg-Marquardt. The method's feasibility and robustness are verified by simulating and real experiments. The proposed algorithm has been successfully applied to a self-development 3D measuring system (TN 3DOMS) of Mianyang Saint Buffalo Technologies Limited Company.