基于机器视觉的车窗按键缺陷检测研究  

Defect Detection of Window Buttons Based on Machine Vision

作　　者：刘伟[1] 王凯[1] 徐洋 胡永超 查长军[1] LIU Wei;WANG Kai;XU Yang;HU Yongchao;ZHA Changjun(School of Advanced manufacturing Engineering,Hefei University,Hefei 230601,China)

机构地区：[1]合肥学院先进制造工程学院,合肥230601

出　　处：《合肥学院学报（综合版）》2022年第5期67-72,共6页Journal of Hefei University：Comprehensive ED

基　　金：安徽省自然科学基金“基于超材料传感增强和深度学习的太赫兹光谱植物油品质快速检测研究”(2008085MC96);安徽省科技重大专项“谷物功能性包装智能成套装备研发及产业化”(202003a06020022)资助。

摘　　要：为解决人工对车窗按键检测过程中速度慢,精度差,成本高的问题,提出了一种基于机器视觉的车窗按键缺陷检测系统。该系统运用BOLB分析的方法,确定好模板,通过模板匹配的算法,进行快速定位,并对待检测图像使用改进的方法分别进行灰度值缩放和去噪,最后运用灰度值差异的方法进行缺陷检测。实验结果表明,该方案具有很好的鲁棒性并且运行速度快,可以达到很好的效果。In order to solve the problems of slow speed,poor precision and high cost in the process of manual detection of window buttons,a window button defect detection system based on machine vision is proposed.The system uses the BOLB analysis method to determine the template,uses the template matching algorithm to quickly locate,and uses the improved method to scale and de-noise the gray value of the image to be detected.Finally,the gray value difference method is used to defect.The experimental results show that the scheme has good robustness and fast running speed,and can achieve good results.

关 键 词：机器视觉 缺陷检测 模板匹配 去噪 车窗按键 

分 类 号：F061.5[经济管理—政治经济学]