一种基于机器视觉的快速规则的表面缺陷检测方法

摘要

提出了一种基于机器视觉的产品表面缺陷快速检测和定位的方法.该方法引用Pear-son(皮尔逊)相关系数的概念,通过待检测样本图像和模板图像皮尔逊相关系数阈值可以很好地确定图像中的产品是否有缺陷,再通过待检测图像和模板图像进行差运算得到差影图像,以实现缺陷的定位.实验结果表明本方法与Du-Ming缺陷检测方法相比具有优越性,并且在壁纸表面缺陷检测中得到了较好的效果.%This paper proposes a fast and regulari detection method for defect detection and localization of product surface.Pearson correlation coefficient is used in this method with the correlated pearson coefficient threshold of the sample and test images to detect and locate the product defects in the test image.The difference between the images is called difference image.Experimental results show that the proposed method has the advantage compared with Du-Ming defect detection method and it is able to detect the de-fects on the surface of wallpaper.