郝 勇,温钦华,吴文辉,邱小川.轴承防尘盖缺陷机器视觉自动检测方法研究[J].测控技术,2020,39(1):80-85 |
轴承防尘盖缺陷机器视觉自动检测方法研究 |
Research on Automatic Measurement Method of Bearing Dust Cover Defects Based on Machine Vision |
|
DOI:10.19708/j.ckjs.2020.01.015 |
中文关键词: 机器视觉 图像处理 连通域检测 最小二乘法 图像展开 |
英文关键词:machine vision image processing connected domain detection least squares image expansion |
基金项目:国家自然科学基金项目(21265006);国家质量监督检验检疫总局科技计划项目(2015IK160);江西省自然科学基金项目(2015ZBAB201003) |
|
摘要点击次数: 1596 |
全文下载次数: 428 |
中文摘要: |
轴承防尘盖在装配过程中易出现各种擦伤、压痕等缺陷,从而影响配套主机的精度、可靠性及寿命,急需研发自动化的轴承表面缺陷检测设备来提高我国轴承生产水平。采用自主设计的自动化检测设备,利用机器视觉结合图像处理方法对轴承表面图像进行采集、分析和处理;采用连通域检测和轮廓筛选方法对轴承外轮廓进行提取;采用最小二乘法拟合外轮廓圆得到圆心及外径,结合半径占比分离法实现防尘盖圆环区域的提取;采用改进算法对防尘盖圆环图像进行展开以避免图像展开时字符或缺陷的误分割;采用垂直方向投影法分离字符区和非字符区,通过连通域检测法对两部分是否存在缺陷分别进行判别。实验结果表明,采用自主设计的轴承检测设备结合最小二乘法、OTSU法、改进的图像展开和投影算法,可实现轴承防尘盖缺陷的自动化检测,正确识别率为96.9%。 |
英文摘要: |
Bearing dust covers are easy to appear a variety of scratches,indentations and other defects during assembly process,which will affect the accuracy,reliability and life expectancy of mechanical host.It is necessary to develop automatic bearing surface defects detection equipment to improve the bearing production level of China.The machine vision combined with image processing methods were used for acquisition,analysis and processing of bearing surface image by using self-designed automation equipment.Connected domain detection and contour screening methods were used to extract the outer contour of the bearing image.The least squares method was used to fit the outer contour circle to obtain the circle center and radius,and the bearing dust cover region was separated according to the ratio of radius.An improved algorithm for expanding dust cover ring image was proposed to avoid the false segmentation of characters or defects.The vertical projection was used to separate the character region and the non-character region,and the connected domain detection method was used to identify the existence of defects in two regions separately.The experimental results show that the self-designed bearing detection equipment combined with least squares method,OTSU method,the improved image expansion algorithm and projection algorithm can realize the automatic detection of the defects in the bearing dust cover with the correct recognition rate of 96.9%. |
查看全文 查看/发表评论 下载PDF阅读器 |
关闭 |