关键词: 工件特征, 三维点云, 法向量估计与调整, 曲面重建

Abstract: Surface reconstruction is the main way to realize industrial reverse and measurement.To improve the reconstruction accuracy of workpiece,a surface reconstruction optimization algorithm was designed according to the characteristics of different feature types.Principal Components Analysis (PCA) method was used to solve the normal vector and curvature of point cloud,and a breadth first search algorithm was proposed to adjust the normal vector direction of point cloud.The sharp features and flat features were distinguished according to the curvature.The sharp features were reconstructed by greedy triangle algorithm and the flat features were reconstructed by Poisson algorithm.This two parts were combined to get the final surface model.Experiments showed that the proposed algorithm was more accurate than greedy triangle algorithm and Poisson algorithm.

Key words: workpiece feature, 3D point cloud, estimation and adjustment of normals, surface reconstruction