Defect detection method based on point cloud information

Abstract

The invention discloses a defect detection method based on point cloud information, and the method comprises the steps: carrying out the data processing of point cloud data of the surface of a detected object, and obtaining a normal vector and a main curvature of each point; segmenting the point cloud data into n local point clouds according to the positions of the coordinates of the point cloud data; calculating the feature number of one of the local point clouds; marking defect points or normal points according to the feature number; repeating until all point cloud data is traversed; according to the marked types of all the points, whether defects exist on the surface of the measured object or not and the positions of the defects are obtained; according to the method, the point cloud data is segmented to improve the defect detection accuracy; multiple times of segmentation are designed, defect points are screened repeatedly, and false detection is effectively prevented; the method issuitable for checking whether deformation defects such as bumps and pits exist on the surface of a product or not, and real-time defect positioning can be carried out.

Description

technical field [0001] The invention relates to the field of defect detection, in particular to a defect detection method based on point cloud information. Background technique [0002] The identification and positioning of product surface defects is a very important link in the product manufacturing process. There are two traditional detection methods: manual detection and image recognition; among them, manual detection is realized by visual and touch methods, which has long detection time, The detection efficiency is low and the detection results are easily affected by the level of workers. This method is generally a random inspection. Once a missed inspection occurs, the product containing defects will enter the next production link, which will affect the product quality. It can be seen that the manual inspection method is difficult to meet modern The needs of industrial production testing. Image recognition is a method of defect analysis based on image contrast. This me...

AI Technical Summary

Problems solved by technology

[0002] The identification and positioning of product surface defects is a very important link in the product manufacturing process. There are two traditional detection methods: manual detection and image recognition; among them, manual detection is realized by visual and touch methods, which has long detection time, The detection efficiency is low and the detection results are easily affected by the level of workers. This method is generally a random inspection. Once a missed inspection occurs, the product containing defects will enter the next production link, which will affect the product quality. It can be seen that the manual inspection method is difficult to meet modern The needs of industrial production testing

Image recognition is a method of defect analysis based on image contrast. This method has high requirements for the environment of the measured object, and the detection sensitivity is low when there are deformation defects such as bulges and pits on the surface of the product.

Therefore, for products with deformation defects such as bulges and pits on the surface, the existing methods cannot provide effective automatic detection methods

Images