基于图像处理分析的真空泵刮片安装缺陷检测  

Defect Detection of Vacuum Pump Blade Installation Based on Image Processing and Analysis

作　　者：邓建波 DENG Jian-bo(Aleph Electronics(Shenzhen)Co.,Ltd.,Shenzhen 518126 China)

机构地区：[1]艾礼富电子(深圳)有限公司,广东深圳518126

出　　处：《自动化技术与应用》2024年第1期44-48,共5页Techniques of Automation and Applications

摘　　要：真空泵刮片是否存在缺陷以及是否正确装入石墨转子,对真空泵的性能影响很大。在真空泵制造装配过程中,采用机器视觉自动判定石墨刮片装入转子后的状态。由于刮片装入转子后其整体的轮廓具有随机性,无法采用轮廓分析进行缺陷检测;刮片及转子成像灰度值不均匀、刮片与背景灰度值对比差,也无法采用传统的模式匹配进行缺陷检测。在借助NI Vision Assistant,依次进行图像预处理提取刮片轮廓后再采用灰度匹配取得良好效果。经过试样结果表明:该方法缺陷识别精度高、结果判定稳定,实现了刮片缺陷检测在工厂现场的应用。Whether there are defects in the vacuum pump blade and whether the blade is correctly installed in the graphite rotor has a great impact on the performance of the vacuum pump.Therefore,in the process of manufacturing and assembling the vacuum pump,machine vision is used to automatically determine the state of the graphite blade after it is installed in the rotor.Because of the randomness of the overall contour after the blades are installed into the rotor,the contour analysis can not be used for defect detection;the gray value of blade and rotor imaging is uneven and the contrast between blade and background gray value is poor,so the traditional pattern matching can not be used for defect detection.With the help of NI Vision Assistant,image preprocessing is carried out successively to extract the blade contour,and then gray matching is used to achieve good results.The sample results show that this method has high defect identification accuracy and stable result judgment,and realizes the application of blade defect detection in factory.

关 键 词：真空泵刮片 机器视觉 缺陷检测 灰度匹配 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置] TP391[自动化与计算机技术—控制科学与工程]