基于分块模板匹配的印刷品文字缺陷图像检测方法  

Image detection method for text defects in printed matter based on block template matching

作　　者：王旭 周海波[1,2] 王军豪 窦富超 任凤雷 WANG Xu;ZHOU Haibo;WANG Junhao;DOU Fuchao;REN Fenglei(Tianjin Key Laboratory for Advanced Mechatronic System Design and Intelligent Control,Tianjin University of Technology,Tianjin 300384,China;National Demonstration Center for Experimental Mechanical and Electrical Engineering Education,Tianjin University of Technology,Tianjin 300384,China)

机构地区：[1]天津理工大学天津市先进机电系统设计与智能控制重点实验室,天津300384 [2]天津理工大学机电工程国家级实验教学示范中心,天津300384

出　　处：《天津理工大学学报》2025年第1期16-22,共7页Journal of Tianjin University of Technology

基　　金：国家重点研发计划项目(2018YFB1308900)。

摘　　要：随着自动化生产的飞速发展,印刷品缺陷检测逐步成为提高印刷质量的重要环节之一。在印刷品文字缺陷检测中,由于受纸张表面柔性等因素影响,存在局部模板匹配精度偏低的问题,导致误检率偏高,为进一步提高其检测精度,文中提出一种基于分块模板匹配的印刷品文字缺陷图像检测方法。该方法将模板图像作为检测标准,利用霍夫直线法检测印刷品裁切线,完成扫描图像整体位姿与大小矫正,并采用分块模板匹配法,完成模板图像与扫描图像的局部高精度匹配,随后经图像正负差分与图像融合处理,获得漏印、墨斑及墨色不均等缺陷图像,最后进行连通域分析去除伪缺陷,进而实现印刷品文字缺陷图像检测。经实验验证,该方法能够较好满足检测要求,在印刷品行业具有一定的应用价值。Printing defect detection has gradually become one of the important links to improve printing quality with the rapid development of automated production.In the detection of printed text defects,due to the influence factors such as the flexibility of the raw paper,there is a problem of low local template matching accuracy,resulting in high false detection rate.In order to further improve its detection accuracy,this paper proposes a printed text defect image detection method based on the block template matching.This method takes the template image as the detection standard,uses the HoughLines to detect the crop line of the printed matter,completes the overall posture and size correction of the scanned image,and adopts the block template matching method to complete the local high-precision matching between the template image and the scanned image,and then obtains the defect image such as missing print,ink spot and ink color uneven by positive and negative difference and fusion processing of the image,and finally performs the connection domain analysis to remove the pseudo-defect,realizing the image detection of the text defect of the printed matter.Experiments have verified that this method can meet the detection requirements well and has certain application value in the printed matter industry.

关 键 词：印刷质量 模板图像 扫描图像 分块模板匹配 缺陷检测 

分 类 号：TH122[机械工程—机械设计及理论]