一种基于螺纹点云的缺陷提取方法

A Defect Extraction Method Based on 3D Thread Point Cloud

针对螺纹表面凹坑缺陷,结合双目三维扫描仪的数据采集特点与螺纹表面特性,提出基于三维螺纹点云缺陷检测的方法。该方法首先利用法向量提取算法,将缺陷部分提取出来,同时将轮廓点云一并提出。然后采用基于密度的DBSCAN聚类算法去除法向量提取的轮廓点云,且只保留凹坑缺陷点云。最后采用最小包围盒算法对缺陷点云参数化,测算其长度、宽度以及深度,实现了表面缺陷的数字化测量。

Aiming at the defect of thread surface pits, combined with the data acquisition characteristics of binocular 3D scanner and thread surface characteristics, a defect detection method based on 3D thread point cloud is proposed.Firstly, the normal vector extraction algorithm is used to extract the defect part, but the contour point cloud is also proposed.Then DBSCAN clustering algorithm based on density is used to remove the contour point cloud extracted by normal vector, and only the pit defect point cloud is retained.Finally, the minimum bounding box algorithm is used to parameterize the defect point cloud, calculate its length, width and depth, and realize the digital measurement of surface defects.