基于FCM算法的套印误差检测  

Overprinting Error Detection Based on FCM Algorithm

作　　者：王煜川 张勇斌 付秀丽[2] WANG Yuchuan;ZHANG Yongbin;FU Xiuli(School of Machanical and Electrical Engineering,Beijing Institute of Graphic Communication,Beijing 102600,China;School of Machanical and Electrical Engineering,Beijing Institute of Petrochemical Technology,Beijing 102617,China)

机构地区：[1]北京印刷学院机电工程学院,北京102600 [2]北京石油化工学院信息工程学院,北京102617

出　　处：《北京印刷学院学报》2023年第12期9-14,共6页Journal of Beijing Institute of Graphic Communication

基　　金：中国高校产学研创新基金项目(2021LDA06007)研究成果。

摘　　要：为了降低四色重叠影响导致的图像色彩分割困难,提高印刷品套印误差检测的精度,提出了一种基于FCM聚类算法的套印误差检测方式。首先,将采集的印刷品套印标识符图像进行灰度化处理、高斯滤波降噪处理以及CMYK色彩空间转换并分离出四种色彩通道的图片;然后,分别对四种色彩通道图片进行FCM算法聚类;最后,对聚类后的图片进行固定阈值分割,并对分割的图片依次做图像差运算,分别提取出四个通道中C、M、Y、K四种颜色的部分,进行误差分析。对比显示测量误差均≤0.2mm,符合国家对一般印刷品套印精度检测误差不高于0.2mm的要求;该算法与应用广泛的传统大津法对比,测量误差更小,精度更高,提升了套印误差检测水平。In order to reduce the difficulty of image color segmentation caused by four color overlap and improve the accuracy of printing overprint error detection,a overprint error detection method based on FCM clustering algorithm is proposed.Firstly,the collected printed matter overprint identifier images are subjected to grayscale processing,Gaussian filtering denoising,and CMYK color space conversion to separate images of four color channels.Then,perform FCM algorithm clustering on the four color channel images separately.Finally,fixed threshold segmentation is performed on the clustered images,and image difference operations are performed on the segmented images in order to extract the parts of C,M,Y,and K colors from the four channels.Error analysis is performed.The comparison shows that the measurement error is≤0.2mm,which meets the national requirement for the detection error of overprint accuracy of general print not to exceed 0.2mm.Compared with the widely used traditional Otsu method,this algorithm has smaller measurement errors and higher accuracy,improving the level of overprint error detection.

关 键 词：套印误差 FCM算法 图像分割 

分 类 号：TS826[轻工技术与工程]