基于机器视觉的金属垫片表面缺陷检测

Metal gasket surface defect detection based on machine vision

针对人工目检方法在色差垫片表面缺陷检测上存在检测速度慢、漏检和误检等问题,以OpenCV开发的柔性视觉检测系统为载体,提出了一种基于机器视觉的金属垫片表面缺陷检测方法。在被检测垫片图像背景加入有色噪声,根据每个垫片图像的颜色分布自动选取一个最佳阈值,对图像进行阈值处理,自适应分割出金属垫片的表面缺陷,从而精确地检测出存在色差垫片的表面缺陷。验证结果表明:该方法能够完成对色差垫片的自动化在线检测,有效提高检测精度和速度。

In order to solve the problems of slow detection speed,missing detection and misdetection of color difference gasket surface defects by manual visual inspection method,this work proposed a metal gasket surface defect detection method based on machine vision by using the flexible visual inspection system developed by OpenCV.Firstly,colored noise was added into the background of the detected gasket image.Then,according to the color distribution of each gasket image,an optimal threshold value is automatically selected.Finally,the image is threshold processing,and the surface defects of metal gasket are segmented adaptively,so as to accurately detect the surface defects of gasket with color difference.The experimental results show that this method can complete the automatic online detection of color difference gaskets,and effectively improve the detection accuracy and speed.