Machine Vision Based on Pipe Joint Surface Defect Detection and Identification

Abstract

In the production process, pipe joints often have surface defects such as cracks, pits, bumps, etc., and the manual detection efficiency is low and the cost is high. This paper presents a machine vision based method for detecting surface defects of pipe joints. The preprocessing algorithm, image initial detection algorithm, image segmentation algorithm and feature extraction algorithm of tube joint surface defect image are studied and BP neural network classifier is designed. Finally, the designed classifier was tested by 300 samples of cracks, pits and bump defects on the surface of the pipe joint. The experimental results show that the recognition rate of the classifier reaches 95.5%, which can better meet the surface of the pipe joint. Defect detection requirements.