Complex profile point cloud normal feature clustering hierarchical estimation method

The invention provides a point cloud normal estimation method based on feature region clustering and grading in order to improve the accuracy of a normal estimation result of a complex profile sampling point cloud containing edges, sharp corners and other features, and belongs to the technical field of product reverse engineering. Performing clustering analysis on the point cloud according to thelocal flatness of the curved surface and the Bayesian information criterion, sequentially dividing the point cloud into flat, feature edge, edge sharp corner and other areas, identifying the feature type of the sample point, iterating the normal estimation result of the sample point in the flat area and propagating the normal estimation result to the adjacent feature area step by step, and obtaining a clustering result; wherein the normal estimation result of the characteristic sample point is consistent with the normal of the adjacent flat area sample point. The method can accurately estimatethe normal of the sample