机器视觉在轧辊表面缺陷检测系统中的应用  

Application of Machine Vision in Roll Surface Defect Detection System

作　　者：石炜[1] 杨锦春 刘松 陈禹州 SHI Wei;YANG Jinchun;LIU Song;CHEN Yuzhou(School of Mechanical Engineering,Inner Mongolia University of Science and Technology,Inner Mongolia Baotou 014010,China)

机构地区：[1]内蒙古科技大学机械工程学院,内蒙古包头014010

出　　处：《机械设计与制造》2025年第2期308-311,共4页Machinery Design & Manufacture

基　　金：2018年内蒙古自治区自然科学基金项目(2018LH05024);2018年内蒙古自治区高等学校科学技术研究项目(NJZY18149)。

摘　　要：轧辊长时间使用表面会出现很多缺陷,为在修复前进行轧辊的缺陷检测,这里提出一种基于机器视觉的轧辊表面缺陷检测方法。利用工业相机获取轧辊表面图像;使用中值滤波算法去除图像中的干扰信息并使用拉普拉斯方法对图像进行锐化处理,增强轧辊表面缺陷区域的边缘信息;然后采用Otsu最佳阈值分割算法对图像进行分割,使图像更容易被理解和分析;最后,使用模板匹配技术进行缺陷识别;对裂纹、腐蚀、辊印这三种缺陷的识别率分别为98.75%、92%、93.33%。该检测系统表明对于腐蚀、裂纹、辊印这3大类轧辊缺陷情况有较好的检测效果。In order to detect the defects of rolls before repair,a method of surface defect detection based on machine vision was pro⁃posed in this paper.The image of the roll surface was obtained by an industrial camera.Median filtering algorithm was used to re⁃move the interference information in the image and Laplace method was used to sharpen the image to enhance the edge informa⁃tion of the roller surface defect area.Then the Otsu optimal threshold segmentation algorithm was used to segment the image,which makes the image easier to understand and analyze.Finally,the template matching technique was used to identify defects.The recognition rates of crack,corrosion and roll printing were 98.75%,92%and 93.33%,respectively.The detection system showed that it had good detection effect on three kinds of roll defects,namely corrosion,crack and roll imprint.

关 键 词：机器视觉 轧辊 表面缺陷 检测系统 

分 类 号：TH16[机械工程—机械制造及自动化] TP391.4[自动化与计算机技术—计算机应用技术]