基于计算机视觉的滚子端面自动检测系统  

Automatic Detection Technology for Roller End Based on Computer Vision

作　　者：翟志永[1] 陈永清[2] 陈廉清[2] ZHAI Zhiyong;CHEN Yongqing;CHEN Lianqing(Mechatronical&Electrical Engineering Institute,Ningbo Polytechnic,Ningbo,Zhejiang 315802,China;Robotics Institute,Ningbo University of Technology,Ningbo,Zhejiang 315211,China)

机构地区：[1]宁波职业技术学院机电工程学院,浙江宁波315800 [2]宁波工程学院机器人学院,浙江宁波315211

出　　处：《轻工机械》2021年第2期74-76,81,共4页Light Industry Machinery

摘　　要：针对轴承滚子端面目检损害工人视力以及误检率高的问题,课题组基于计算机视觉检测技术设计了轴承滚子端面自动检测系统。首先,设计了图像采集系统采集滚子端面图像,分析滚子端面缺陷类型的特点,通过对图像的滤波处理、最大类间方差阈值法二值化分割,提取了缺陷特征;然后,分别设定相应的感兴趣检测区域(ROI),定义检测目标的最大像素数作为判别缺陷特征的依据;最后,应用LabVIEW软件开发了计算机视觉自动检测软件系统。实践表明该系统能很好地实现滚子双端面缺陷的自动检测,能满足准确、高效、客观的实时检测要求。Aiming at the problem that the visual inspection of bearing roller end damages the eyesight of workers and the high false detection rate,the automatic detection system for roller end based on computer vision inspection technology was proposed.The image acquisition system was designed to collect the image of roller end.The characteristics of roller end defects were analyzed,the defect features were extracted by filtering the image and the binarization processing of maximum inter class variance threshold method.And then the corresponding ROI(region of interest)was set to define the maximum number of pixels of the detection target as the basis to distinguish the defect features.The computer vision automatic detection system was developed by using LabVIEW software.The practice shows that the system can realize the automatic detection of roller double end defects and meet the requirements of accurate,efficient and objective realtime detection.

关 键 词：计算机视觉 轴承滚子端面 缺陷特征 最大类间方差阈值法 感兴趣检测区域 LABVIEW软件 

分 类 号：TH165.4[机械工程—机械制造及自动化] TP391[自动化与计算机技术—计算机应用技术]