基于人眼视觉特性的印刷品缺陷形态学检测方法  

Morphological Scheme for Detection of Printing Defect Based on Human Visual System

作　　者：杨欧[1] 彭艳[1] 戴光智[1] 

机构地区：[1]深圳职业技术学院计算机工程学院,深圳578055

出　　处：《现代计算机（中旬刊）》2012年第2期31-34,44,共5页Modern Computer

基　　金：广东省自然科学基金(No.07006479);深圳职业技术学院青年创新基金(No.2209k3010015)

摘　　要：提出一种基于人眼视觉的印刷品缺陷检测方法,方法根据印刷品缺陷检测的特点,采用减影算法检测印刷品图像的缺陷。提出动态灰度阈值的减影改进算法,此方法符合人眼对于不同灰度差别,具有不同的鉴别能力的特点。结合形态学预处理建模算法,在获得准确缺陷的情况下,能基本消除轮廓伪影,得到真实的印刷缺陷。Presents a print defect detection method based on human visual. This method uses the subtrac- tion algorithm to detect the defects of the printed image according to the characteristics of the print defect detection. Proposes the subtraction improved algorithm of dynamic gray-scale threshold, and this approach is consistent with the human eye, has the ability to identify the characteristics of different gray. Based on the morphology pretreatment modeling algorithm, and in the case of access to accurate defect, can basically eliminate the contour artifacts, and get the real printing defects.

关 键 词：灰度形态学 印刷品缺陷检测 人眼视觉特性 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]