摘要
为了满足轴承内外径尺寸检测的高精度需求,解决人工检测误差大以及效率低的问题,提出了一种基于图优化的轴承尺寸检测优化方法。首先,对灰度图像采用双边滤波去除噪声,通过形态学操作排除图像中的小型黑洞,使用图像分割提取图像中感兴趣的目标;其次,运用Canny边缘检测算子对图像边缘进行粗定位,再用Zernike-Otsu法获取图像的亚像素边缘坐标;最后,把亚像素边缘坐标导入g2o优化模型进行曲线拟合得到轴承的内外径尺寸。实验结果表明,利用该方法进行轴承内外径检测的误差小于0.015 mm,能够满足工业现场的检测需求。
To meet the high-precision requirements of size inspection,and to solve the problems of large manual inspection errors and low efficiency,a bearing size inspection optimization method based on graph optimization is proposed.First,use bilateral filtering to remove noise on the grayscale image,eliminate small black holes in the image through morphological operation,and use image segmentation to extract the target of interest in the image;second,use Canny edge detection operator to roughly locate the image edge,and then Use the Zernike-Otsu method to obtain the sub-pixel edge coordinates of the image;finally,import the sub-pixel edge coordinates into the g2o optimization model and perform curve fitting to obtain the inner and outer diameter dimensions of the bearing.The experimental results show that the internal error of the bearing detected by this method is less than 0.015 mm,which can meet the detection requirements of industrial sites.
作者
陈甦欣
韩暑
涂德江
晏文彬
CHEN Su-xin;HAN Shu;TU De-jiang;YAN Wen-bin(School of Mechanical Engineering, Hefei University of Technology, Hefei 230009,China)
出处
《组合机床与自动化加工技术》
北大核心
2021年第2期103-106,共4页
Modular Machine Tool & Automatic Manufacturing Technique
基金
国家产业技术基础公共服务平台项目(2019-00899-2-1)。
关键词
尺寸检测
图像分割
亚像素
图优化
轴承
dimension detection
image segmentation
sub-pixel
graph optimization
bearing