用于漆面缺陷识别检测的数字图像处理技术系统设计
Design of A Digital Image Processing Technology System for Identifying and Detecting Paint Surface Defects
史章潞 1韩玉振 1彭一真 1徐文江 1董洪伟 1王利军1
作者信息
- 1. 一汽-大众汽车有限公司成都分公司,成都 610199
- 折叠
摘要
自动喷漆过程后车身表面的小缺陷是目前汽车行业质量控制面临的最大问题之一.利用图像处理技术可以实现一种新的检测方法,并确保整个区域得到检查.处理方法是基于局部定向模糊方法的图像背景提取方法和修改后的图像对比度增强,从而能够检测整个照明区域的缺陷.此外,还可以使用多级结构进行多次处理,该系统已四川成都的大众工厂的生产线上试运行.这里将展示对性能的完整分析,以及证明稳健性和可靠性的几个测试.
Abstract
Small defects on the surface of the bodywork after an automated painting process are one of the biggest problems facing quality control in the automotive industry today.Image processing is used to realize a new inspection method and to ensure that the en-tire area is inspected.The processing method is i.e.an image background extraction method based on the local directional blurring method and a modified image contrast enhancement,which enables the detection of defects throughout the illuminated area.In addi-tion,the system has been trialed on the production line at the Volkswagen plant in Chengdu,Sichuan,using a multi-stage architecture for multiple processing.A complete analysis of the performance is shown here,as well as several tests to demonstrate robustness and reliability.
关键词
数字图像处理/车身检查/人机协同/自动视觉检查Key words
digital image processing/vehicle inspection/human-machine collaboration/automatic visual inspection引用本文复制引用
出版年
2024
NETL
NSTL
万方数据