目的 解决包装材料高速生产中印刷缺陷与防伪特征难以同步在线判定、质量事件无法及时驱动生产调度的问题,为印刷装备提供兼具高精度检测与实时调度能力的工程化解决方案,提升生产质量与交付效率。方法 构建基于改进YOLOv5框架的双任务并行检测模型,融合注意力机制、小目标检测层与孪生比对,并与事件驱动的动态调度闭环集成,打通算法结果到PLC执行指令的毫秒级转化通道。结果 实验结果显示,总体平均精度（阈值区间0.50⁓0.95）为0.694,小目标平均精度为0.519;在误报率为1%时的检出率为0.939,归一化互相关均值为0.868、光谱角为6.07度;端到端时延9.9 ms、帧率为100.8帧/s;平均订单延迟降至1.73 h、班均停机时长降至17.9 min。结论 模型实现了印刷缺陷与防伪特征的同框并行有效检测,事件驱动调度机制有效解决了质量事件与生产调度的协同问题,研究为印刷装备在线质量控制与防伪追溯提供可落地的工程化路径。
Abstract
The work aims to solve the problem of difficulty in synchronizing online judgment of printing defects and anti-counterfeiting features in high-speed production of packaging materials, and the inability to drive production scheduling in a timely manner due to quality events, so as to provide an engineering solution for printing equipment that combines high-precision detection and real-time scheduling capabilities, and to improve production quality and delivery efficiency. A dual task parallel detection model was built based on the improved YOLOv5 framework. Through integrating the attention mechanism, the small object detection layer, and twin comparison, and integrating with event driven dynamic scheduling closed-loop, a millisecond level conversion channel from algorithm results to PLC execution instructions was constructed. The experimental results showed that the overall average accuracy (threshold range 0.50-0.95) was 0.694, and the average accuracy of small targets was 0.519; The detection rate at a false positive rate of 1% was 0.939, with a normalized cross-correlation mean of 0.868 and a spectral angle of 6.07 degrees; end to end latency of 9.9 milliseconds and frame rate of 100.8 frames per second; The average order delay was reduced to 1.73 hours, and the average downtime per shift was reduced to 17.9 minutes. In conclusion, the model achieves high-precision detection of printing defects and anti-counterfeiting features in parallel within the same frame. The event driven scheduling mechanism effectively solves the coordination problem between quality events and production scheduling. The research provides a practical engineering path and method basis for online quality control and anti-counterfeiting traceability of printing equipment.
关键词
印刷缺陷检测 /
防伪识别 /
目标检测 /
事件驱动调度 /
质量控制
Key words
printing defect detection /
anti-counterfeiting identification /
object detection /
event driven scheduling /
quality control
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] 李博宇, 杨文翰, 杨乐, 等. 基于模板匹配的肥料外包装印刷缺陷检测研究[J]. 印刷与数字媒体技术研究, 2023(2): 39-49.
LI B Y, YANG W H, YANG L, et al.Research on Defect Detection of Fertilizer Packaging Based on Template Matching[J]. Printing and Digital Media Technology Study, 2023(2): 39-49.
[2] 周著国, 鲁玉军, 吕利叶. 基于改进YOLOv5s的印刷电路板缺陷检测算法[J]. 浙江大学学报(工学版), 2025, 59(8): 1608-1616.
ZHOU Z G, LU Y J, LYU L Y.Improved YOLOv5s-Based Algorithm for Printed Circuit Board Defect Detection[J]. Journal of Zhejiang University (Engineering Science), 2025, 59(8): 1608-1616.
[3] 孙铁强, 于洪健, 张灿, 等. 基于YOLO提示与ICT-ViT的印刷电路板缺陷高效分割技术研究[J]. 激光与光电子学进展, 2025, 62(12): 185-195.
SUN T Q, YU H J, ZHANG C, et al.Efficient Printed Circuit Board Defect Segmentation Technology Based on YOLO Prompts and ICT-ViT[J]. Laser & Optoelectronics Progress, 2025, 62(12): 185-195.
[4] 范博淦, 王淑青, 陈开元. 基于改进YOLOv8的印刷电路板缺陷检测模型[J]. 现代电子技术, 2025, 48(11): 144-150.
FAN B G, WANG S Q, CHEN K Y.Printed Circuit Board Defect Detection Model Based on Improved YOLOv8[J]. Modern Electronics Technique, 2025, 48(11): 144-150.
[5] 毛坤, 朱学军, 赖惠鸽, 等. 多尺度融合优化的印刷电路板缺陷检测算法[J]. 激光与光电子学进展, 2025, 62(10): 1012004.
MAO K, ZHU X J, LAI H G, et al.Multi-Scale Fusion Optimization Algorithm for Printed Circuit Board Defect Detection[J]. Laser & Optoelectronics Progress, 2025, 62(10): 1012004.
[6] 张鹏涛, 李文峰, 宋强. 基于注意力机制的印刷包装标签文本自动定位检测方法[J]. 包装与食品机械, 2025, 43(3): 80-87.
ZHANG P T, LI W F, SONG Q.Attention Mechanism-Based Automatic Localization and Detection Method for Text in Printed Packaging Labels[J]. Packaging and Food Machinery, 2025, 43(3): 80-87.
[7] 张小雪, 孙帮勇, 谭家海, 等. 印刷品缺陷检测方法综述[J]. 印刷与数字媒体技术研究, 2025(2): 1-10.
ZHANG X X, SUN B Y, TAN J H, et al.A Review of Defect Detection Methods for Printed Products[J]. Printing and Digital Media Technology Study, 2025(2): 1-10.
[8] 李大红, 李伟, 高峰. 人工智能驱动下包装印刷行业产学研协同创新机制构建与实践路径研究[J]. 印刷与数字媒体技术研究, 2025(3): 265-273.
LI D H, LI W, GAO F.Research on Mechanism Construction and Practical Path of Industry-Education-Research Collaborative Innovation in Packaging and Printing Industry Driven by Artificial Intelligence[J]. Printing and Digital Media Technology Study, 2025(3): 265-273.
[9] PETKOV N, IVANOVA M.Printed Circuit Board and Printed Circuit Board Assembly Methods for Testing and Visual Inspection: A Review[J]. Bulletin of Electrical Engineering and Informatics, 2024, 13(4): 2566-2585.
[10] PATPATIYA P, CHAUDHARY K, KAPOOR V.Reverse Manufacturing and 3D Inspection of Mechanical Fasteners Fabricated Using Photopolymer Jetting Technology[J]. MAPAN, 2022, 37(4): 753-763.
[11] 杨奡飞, 续欣莹, 谢刚, 等. 印刷电路板缺陷持续检测与定位方法研究[J]. 智能系统学报, 2025, 20(1): 219-229.
YANG A F, XU X Y, XIE G, et al.Research on Continual Detection and Localization Method for Printed Circuit Board Defect[J]. CAAI Transactions on Intelligent Systems, 2025, 20(1): 219-229.
[12] 沈胤熙, 周海燕, 樊晨龙, 等. 人工智能在刨花板外观质量检测方面的应用[J]. 世界林业研究, 2025, 38(1): 80-84.
SHEN Y X, ZHOU H Y, FAN C L, et al.Application of Artificial Intelligence in Particleboard Surface Quality Detection[J]. World Forestry Research, 2025, 38(1): 80-84.
[13] 李大华, 徐傲, 王笋, 等. 基于改进YOLOv5的印刷电路板缺陷检测[J]. 电子测量技术, 2023, 46(23): 112-119.
LI D H, XU A, WANG S, et al.Printed Circuit Board Defect Detection Based on Improved YOLOv5[J]. Electronic Measurement Technology, 2023, 46(23): 112-119.
[14] JAVAID M, HALEEM A, SINGH R P, et al.Artificial Intelligence Applications for Industry 4.0: A Literature-Based Study[J]. Journal of Industrial Integration and Management, 2022, 7(1): 83-111.
[15] 卞佰成, 陈田, 吴入军, 等. 基于改进YOLOv3的印刷电路板缺陷检测算法[J]. 浙江大学学报(工学版), 2023, 57(4): 735-743.
BIAN B C, CHEN T, WU R J, et al.Improved YOLOv3-Based Defect Detection Algorithm for Printed Circuit Board[J]. Journal of Zhejiang University (Engineering Science), 2023, 57(4): 735-743.
[16] 赵晓, 王建聪, 齐浩天. 基于改进YOLOv8的印刷电路板缺陷检测算法[J]. 计算机工程与设计, 2025, 46(10): 2994-3000.
ZHAO X, WANG J C, QI H T.Improved YOLOv8-Based Algorithm for Printed Circuit Board Defect Detection[J]. Computer Engineering and Design, 2025, 46(10): 2994-3000.
[17] 王海涛, 高玉栋, 侯建新, 等. 基于DCGAN的印刷缺陷检测方法[J]. 哈尔滨理工大学学报, 2021, 26(6): 24-32.
WANG H T, GAO Y D, HOU J X, et al.A Method of Printing Defect Detection Based on DCGAN[J]. Journal of Harbin University of Science and Technology, 2021, 26(6): 24-32.
[18] 王佑芯, 陈斌. 基于深度对比网络的印刷缺陷检测方法[J]. 计算机应用, 2023, 43(1): 250-258.
WANG Y X, CHEN B.Print Defect Detection Method Based on Deep Comparison Network[J]. Journal of Computer Applications, 2023, 43(1): 250-258.
[19] 梁承权, 吕德深, 陆晓. 一种基于改进Faster RCNN的易拉罐印刷缺陷检测方法[J]. 印刷与数字媒体技术研究, 2023(6): 22-29.
LIANG C Q, LYU D S, LU X.A Method for Detecting the Printing Defects of Easy Open Can Based on Improved Faster RCNN[J]. Printing and Digital Media Technology Study, 2023(6): 22-29.
[20] 刘国庆, 方成刚, 黄德军, 等. 基于改进YOLOv5的试剂卡印刷缺陷检测算法[J]. 包装工程, 2023, 44(17): 197-205.
LIU G Q, FANG C G, HUANG D J, et al.Reagent Card Printing Defect Detection Algorithm Based on Improved YOLOv5[J]. Packaging Engineering, 2023, 44(17): 197-205.
PDF(731 KB)
渝公网安备 50010702501716号 渝ICP备15012534号-2
