作者：项新建, 王乐乐, 曾航明

作者单位：浙江科技学院自动化与电气工程学院，浙江 杭州 310023

关键词：推力轴承垫圈;机器视觉;缺陷检测;图像处理;傅里叶变换

摘要：

为实现推力轴承垫圈表面缺陷自动检测，解决检测过程人工效率低、准确性波动大的问题，设计一套推力轴承垫圈表面缺陷检测及分类系统。针对垫圈表面图像背景复杂，干扰较多难以提取缺陷等问题，用最小二乘法圆拟合提取圆环形感兴趣区域，通过傅里叶变换、低通滤波器和傅里叶逆变换对图像卷积滤波从而抑制干扰；然后基于多个形状指标和多个阈值提出缺陷提取分类算法，从而达到检测缺陷和分类的目标。通过实验并分析结果：对单个垫圈检测时间为1064.89 ms，垫圈端面缺陷检测准确率在95.23%以上，满足实际检测要求，有具体使用价值，可为自动化检测推力轴承垫圈的缺陷提供新的方法。

Research of defect detection system for thrust bearing gasket based on machine vision
XIANG Xinjian, WANG Lele, ZENG Hangming
School of Automation and Electrical Engineering, Zhejiang University of Science and Technology, Hangzhou 310023, China
Abstract: In order to achieve the automatic detection surface defects of thrust bearing gasket, solve the problems of low artificial efficiency and large fluctuation of accuracy in the detection process. A surface defects detection and classification system of thrust bearing gaskets is designed. In view of the complex background of the image on the gasket surface and the difficulty in extracting defects due to too much interference, the least square method is used to extract the circular region of interest, and the image convolution was filtered by the Fourier transform, low pass filter and inverse Fourier transform to suppress the interference. Then a defect extraction and classification algorithm based on multiple shape indexes and multiple threshold is proposed, to achieve the goal of defect detection and classification. The results of the experiment and analysis show that the detection time of a single gasket is 1064.89 ms, the detection accuracy of the end surface defect of the gasket is more than 95.23%, which can meet the actual detection requirements. It has specific application value and provides a new method for automatic detection of the defects of the thrust bearing gasket.
Keywords: thrust bearing gasket;machine vision;defect detection;image processing;Fourier transform
2021, 47(2):133-139  收稿日期: 2020-08-07;收到修改稿日期: 2020-09-16
基金项目:
作者简介: 项新建(1964-),男,浙江永康市人,教授,硕士,主要从事人工智能、机器人、物联网理论与技术研究
参考文献
[1] LIU B, YANG Y Q, WANG S S, et al. An automatic system for bearing surface tiny defect detection based on multi-angle illuminations[J]. Optik, 2020, 208: 164517
[2] AMINZADEH M, KURFESS T. Automatic thresholding for defect detection by background histogram mode extents[J]. Journal of Manufacturing Systems, 2015, 37: 83-92
[3] 张松林. 轴承手册[M]. 南昌: 江西科学技术出版社. 2004.
[4] 杜晓辉, 刘霖, 张静, 等. 火花塞端面缺陷自动检测算法设计[J]. 计算机技术与发展, 2019, 29(2): 172-176
[5] 韩志玮, 高美凤. 刹车片表面缺陷的图像检测方法[J]. 应用光学, 2020, 41(3): 538-547
[6] MARTÍNEZ S S, VÁZQUEZ C O, GARCÍA J G, et al. Quality inspection of machined metal parts using an image fusion technique[J]. Measurement, 2017, 111: 374-383
[7] PRAPPACHER N, BULLMANN M, BOHN G, et al. Defect detection on rolling element surface scans using neural image segmentation[J]. Applied Sciences, 2020, 10(9): 3290
[8] 陈琦, 阮鸿雁. 基于机器视觉的滑动轴承缺陷检测系统设计[J]. 组合机床与自动化加工技术, 2017(5): 92-95,99
[9] HEMMATI F, MIRASKARI M, GADALA M S. Application of wavelet packet transform in roller bearing fault detection and life estimation[J]. Journal of Physics: Conference Series, 2018, 1074(1): 012142
[10] 汪凤林, 周扬, 叶绿, 等. 基于机器视觉的飞轮齿圈缺陷和尺寸检测方法[J]. 中国测试, 2019, 46(5): 31-38
[11] 韩亮. 基于机器视觉的轴承内外径尺寸测量方法[J]. 机械制造与自动化, 2020, 49(2): 229-231
[12] GONZALEZ R C, WOODS R E, EDDINS S L. 数字图像处理的MATLAB实现(第2版)[M]. 阮秋琦, 译. 北京: 清华大学出版社, 2013.
[13] LIANG J J, ZHAO J, HAO Y X. An image denoising and enhancement algorithm for inner and outer ring of wavelet bearings based on improved threshold[J]. Journal of Physics: Conference Series, 2018, 1069(1): 012158
[14] 王义文, 屈冠彤, 付鹏强, 等. 基于机器视觉的光栅表面缺陷检测系统[J]. 光电工程, 2016, 43(9): 14