面向机油泵零件关键尺寸的机器视觉测量  

Machine Vision Measurement of Critical Dimensions for Oil Pump Parts

作　　者：杨建西[1] 林海波[1] YANG Jian-xi;LIN Hai-bo(School of Mechanical Engineering,Taizhou Vocational & Technical College, Taizhou Zhejiang 318000, China)

机构地区：[1]台州职业技术学院机电工程学院

出　　处：《组合机床与自动化加工技术》2019年第6期54-57,共4页Modular Machine Tool & Automatic Manufacturing Technique

基　　金：浙江省教育厅科研项目(Y201840438);2019年度台州职业技术学院校级课题(2019ZD02,2019ZD03)

摘　　要：机油泵工作平面的检测精度直接影响其工作稳定性和寿命,针对利用微分算子进行边缘检测存在"提升噪声"缺陷,提出一种基于机器视觉的测量方法。在对背光数字图像进行边缘亚像素边缘提取的基础上,采用Ramer算法对获取的亚像素边缘坐标数据按几何特征进行分段,采用改进最小二乘法,提取有用边缘,迭代拟合零件轮廓,抑制离值点对边缘检测干扰。通过机油泵内外转子的检测结果表明:该方法能够精确提取具有清晰边缘的高精度复杂形状背光图像边缘信息,高效应用于中小机械零件的中心距、圆度等几何量的中高精度(IT5~IT7)测量。The detection accuracy of the working plane of the oil pump directly affects its working stability and life. For the edge detection using differential operator, there is a defect of "lifting noise", and a measurement method based on machine vision is proposed. Based on the edge sub-pixel edge extraction of the backlit digital image, the sub-pixel edge coordinate data obtained by the Ramer algorithm are segmented according to geometric features. The improved least squares method is used to extract useful edges and iteratively fit the contour of the part to suppress The off-point points interfere with edge detection. The experimental results on internal and external rotor of oil pump show that the method is helpful and accurate to extract edge of high-precision and complex shaped backlight images and can be applied to high-precision (IT5~IT7) geometric quantity measurements, such as center distance and roundness of small workpiece.

关 键 词：机油泵转子 显微测量 曲面拟合 亚像素边缘 Ramer算法 

分 类 号：TH13[机械工程—机械制造及自动化] TG506[金属学及工艺—金属切削加工及机床]