摘要
针对工业流水线中轴承工件人工检测及装配效率低的问题,提岀一种基于高斯加权均值分割的轴承滚子检测和轴承保持架支柱定位方法。首先,根据移位方法将原始图像转变为灰度图像,利用高斯加权均值分割算法对灰度图像进行阈值分割,以解决二值图像边界较为模糊以及不完整的问题,从而提高边界轮廓检测精度,并采用8邻域边界跟踪算法以提取出二值图像的边界轮廓;在此基础上,根据目标轮廓的几何性质设计轮廓筛选策略对非目标轮廓进行筛选,实现轴承滚子的准确检测,推导出目标轮廓的重心坐标,实现轴承保持架的准确定位;最后,通过轴承滚子检测和轴承保持架定位两个实验,发现所提方法比传统方法检测率提高了12.66%,轴承保持架支柱定位点处的圆拟合误差在水平方向降低0.0657 cm,垂直方向降低0.1189 cm(相对轴承外圈的圆心坐标),且在轴承保持架支柱定位点至轴承外圈圆心的平均距离偏差减少0.0640 c m,从而验证了所提方法的有效性与可行性。
Aiming at the problem of manual detection and assembly efficiency of bearing workpieces in industrial assembly line,a method of bearing roller detection and bearing cage strut positioning based on Gaussian weighted mean segmentation is proposed.Firstly,according to the shift method,the original image is transformed into gray image,and the Gaussian weighted mean segmentation algorithm is used to threshold the gray image to solve the problem that the boundary of the binary image is fuzzy and incomplete,so that improving the accuracy of boundary contour detection.The 8-neighborhood boundary tracking algorithm is used to extract the boundary contour of the binary image.On this basis,according to the geometric properties of the target contour,a contour screening strategy is designed to screen the non-target contour,so as to realize the accurate detection of bearing rollers,derive the center of gravity coordinates of the target contour,and realize the accurate positioning of bearing cages.Finally,through two experiments of bearing roller detection and bearing cage positioning,it is found that the detection rate of the method proposed in this paper is 12.66%higher than that of the traditional method.The circle fitting error at the bearing cage support positioning point is reduced by 0.0657 cm in the horizontal direction and 0.1189 cm in the vertical direction(the circle coordinate is relative to the outer ring of the bearing).The average distance deviation between the support position point of the cage and the center of the bearing outer ring is reduced by 0.0640 cm,which verifies the effectiveness and feasibility of the method proposed in this paper.
作者
魏利胜
丁坤
段志达
Wei Lisheng;Ding Kun;Duan Zhida(School of Electrical Engineering,Anhui Polytechnic University,Wuhu 241000,China)
出处
《电子测量与仪器学报》
CSCD
北大核心
2019年第10期118-127,共10页
Journal of Electronic Measurement and Instrumentation
基金
安徽省自然科学基金(1608085MF146)
安徽工程大学中青年拔尖人才项目(2016BJRC008)资助
关键词
机器视觉
阈值分割
轮廓检测
轴承定位
machine vision
threshold segmentation
contour detection
bearing positioning