小型油泵叶片尺寸高精度测量研究  

Research of High-precision Measurement of Small Oil Pump Blade

作　　者：周晓[1] 谢明棚 ZHOU Xiao;XIE Mingpeng(School of Mechanical and Electronic Engineering,Wuhan University of Technology,Wuhan 430070,China)

机构地区：[1]武汉理工大学机电工程学院,湖北武汉430070

出　　处：《数字制造科学》2022年第3期169-174,共6页

摘　　要：针对小型油泵叶片尺寸的高精度测量要求,提出了一种基于亚像素边缘的尺寸检测方法。根据标定板图像,先对相机畸变分区域校正,建立像素坐标到实际坐标的映射索引;在像距和物距不变的情况下,采集油泵叶片图像,采用Canny算子和Sigmoid函数拟合边缘,获取边缘亚像素坐标,并将其映射到实际坐标中;最后进行最小二乘边缘拟合,计算尺寸参数。试验表明,该方法相比于全局相机标定,检测结果更可靠,绝对精度误差≤15μm,满足油泵叶片的检测精度。To meet the requirement of the dimensional measuring accuracy of the oil pump blade,a dimensional detecting method based on sub-pixel edges is proposed.According to the calibration plate image,the distortion is firstly corrected,then build the mapping index between the actual coordinate and the pixel coordinate.By maintaining the image distance and the object distance,the oil pump blade image is acquired.The edge pixels of the blade are fitted with Canny operator and Sigmoid function to obtain the edge sub-pixel coordinates,in order to integrate them into actual coordinates.Finally,performing the least-squares edge fitting process to calculate the dimensional parameters.The testing results show that the method is more reliable if comparing with global camera calibration.The absolute accuracy error is below 15 microns,which meets the detection accuracy of oil pump blade.

关 键 词：油泵叶片 尺寸测量 图像校正 亚像素 高精度 

分 类 号：TG801[金属学及工艺—公差测量技术] TP391.41[自动化与计算机技术—计算机应用技术]