基于机器视觉算法的轴承套端面缺陷检测研究  

Defect detection algorithm for bearing end face based on machine vision technology

作　　者：骆腾斌 陈硕[1] 赵紫阳 连青惠 

机构地区：[1]福州大学机械工程及自动化学院,福建福州350116 [2]福建省永安轴承有限责任公司,福建永安366000

出　　处：《机电工程》2018年第2期148-152,共5页Journal of Mechanical & Electrical Engineering

摘　　要：针对目前人工检测轴承套端面缺陷存在的效率低、人为因素影响大等诸多问题,提出了一种应用机器视觉技术实现轴承套端面缺陷检测的方法。首先,对采集到的图像进行了平滑处理,并运用自适应阈值的Canny算子完成了边缘检测;其次,利用最小二乘法拟合了轮廓圆,从而提取出了轴承端面的圆环区域;然后,通过Otsu算法计算出了圆环区域的最佳阈值,实现了阈值分割;最后,通过提取各连通域的特征来检测和判别缺陷。实验结果表明:该方法能有效地检测出轴承套端面存在的缺陷,且误检率低于3%,漏检率低于1%,检测时间不超过50 ms,可满足在线检测要求。Aiming at the problems of low efficiency and high influence of human factors in manual inspection of bearing end face defects,a method was proposed to detect the defects of bearing end surface by machine vision technology. Firstly,the image was smoothed and the edge detection was done by using the Canny operator of adaptive threshold. Secondly,the contour circle was fitted by the least square method to extract the annular area of the bearing end face. Thirdly,the optimal threshold of the ring region was calculated by Otsu algorithm,so that the threshold segmentation can be realized. Finally,the defects were detected and discriminated by extracting the characteristics of the connected domains. The experimental results indicate that the method can effectively detect the defects. The false detection rate is less than 3% and the undetected rate is less than 1%. The detection time is less than 50 ms. The proposed algorithm can meet the requirements of on-line inspection.

关 键 词：轴承 表面检测 OTSU算法 机器视觉 

分 类 号：TH133.3[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]