Clip Feature Extraction Method Based on 3D Point Cloud

The detection of clip defects plays a decisive role in the safety of important infrastructure, but the clip parameters reach the micron level, and the accuracy and efficiency of traditional online detection methods cannot be balanced. In this paper, based on the three-dimensional measurement principle of linear structured light, a set of automatic acquisition system of three-dimensional point cloud data on the inner surface of the clip is constructed. Its horizontal resolution is 0.0635mm, and the elevation resolution is 0.00195mm. The collected data can completely characterize the characteristics of the inner surface of the clip. The profile data is used to restore the thread tooth with high accuracy, and its characteristic parameters are extracted. By comparing the design requirements of parameters, single section marks defects, and then all sections use fast connected domain analysis to mark point clouds to determine clip defects. The experiment shows that the extraction error of thread feature parameters is 0.04mm, the defect recognition rate is 98%, and the detection speed of a pair of clips is within 0.45s.