一种基于机器视觉的产品外包装印刷缺陷检测技术  

A Machine Vision Based Printing Defect Detection Technology for Product Packaging

作　　者：汪以歆 余光辉 徐镪 范洪辉 WANG Yi-xin;YU Guang-hui;XU Qiang;FAN Hong-hui(School of Computer Engineering,Jiangsu University of Technology,Changzhou 213001,China)

机构地区：[1]江苏理工学院计算机工程学院

出　　处：《江苏理工学院学报》2019年第4期7-14,共8页Journal of Jiangsu University of Technology

基　　金：国家自然科学基金“基于改进稀疏表示与多特征融合的道路监控视频中车辆信息感知”（61806088）;常州市科技支撑计划（社会发展）“安防监控视频数据智能分析关键技术研究与应用”（CE20175026）;江苏省研究生实践创新计划项目“基于机器视觉的小型机械零部件定量检测系统研究”（SJCX18_1047)

摘　　要：为了提高产品外包装印刷缺陷检测的效率、稳定性和灵活性,提出了一种基于机器视觉的产品外包装印刷缺陷检测技术。引入基于机器视觉的比对检测算法,改进传统的产品外包装印刷缺陷检测方法。试验表明,基于机器视觉的产品外包装印刷缺陷检测技术在不同的试验条件下都能稳定应用,可以快速精确地检测出印刷缺陷。该技术有效地解决了人工抽样检测结果误差大、主观性强、检测速度慢等问题,具有更快的效率、更好的稳定性和处理的灵活性。In order to improve the efficiency, stability and flexibility of the printing defect detection of product packaging,a new technology based on machine vision was proposed in this paper. The traditional method of printing defect detection for external packaging of products was modified,and the comparison detection algorithm based on machine vision was introduced to improve the efficiency,stability and flexibility of the detection. Experimental results show that the machine vision based printing defect detection technology has strong stability in different experimental environments,and can quickly and accurately detect the printing defects. The proposed technology effectively solves the problems of large error of sampling detection results, strong subjectivity of manual detection and slow detection speed, which have faster execution efficiency, better stability and flexibility of processing compared with traditional detection methods.

关 键 词：机器视觉 产品包装 印刷缺陷 缺陷检测 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]