基于机器视觉的沉头孔质量检测方法  

Quality inspection method of counterbore based on machine vision

作　　者：郝博 郭嵩 王婵娟 闫俊伟 HAO Bo;GUO Song;WANG Chanjuan;YAN Junwei(School of Mechanical Engineering and Automation,Northeastern University,Shenyang 110819,CHN;School of Control Engineering,Northeast University at Qinhuangdao,Qinhuangdao 066004,CHN)

机构地区：[1]东北大学机械工程与自动化学院,辽宁沈阳110819 [2]东北大学秦皇岛分校控制工程学院,河北秦皇岛066004

出　　处：《制造技术与机床》2021年第11期100-106,共7页Manufacturing Technology & Machine Tool

基　　金：装备预研领域基金重点项目(61409230125)。

摘　　要：针对沉头孔质量的检测效率低下和检测精度不稳定等问题,提出一种基于机器视觉的沉头孔质量检测方法。该方法使用工业相机采集沉头孔图像,并用高斯滤波等方法处理图像并提取所需特征值;通过建立沉头孔的平行投影计算模型得到其垂直度法矢偏差与锪窝深度的计算公式,结合透视投影计算模型对透视畸变引起的误差进行修正;最后,为验证方法的有效性进行测量试验。试验结果表明,该检测方法测得沉头孔垂直度矢量偏差的误差约为0.04°,锪窝深度的误差约为0.03 mm。Aiming at the problems of low inspection efficiency and unstable inspection accuracy of counterbore quality,a quality inspection method of countersunk hole based on machine vision was proposed.In this method,the image of counterbore is collected by an industrial camera,and the image is processed by Gaussian filtering and the required eigenvalues are extracted.By establishing the mathematical model of parallel projection of counterbore and modifying it with perspective projection,the vertical normal vector deviation and dimple depth of countersunk hole are obtained.Finally,the measurement test is carried out to verify the effectiveness of the method.The test results show that the error of verticality vector deviation is about 0.04°and the error of dimple depth is about 0.03 mm,which can meet the engineering requirements.

关 键 词：机器视觉 沉头孔 垂直度法矢偏差 锪窝深度 

分 类 号：TH878[机械工程—仪器科学与技术]