目的 针对胶印过程中网点变形导致的色彩还原不准确和印刷质量下降问题,提出一种基于机器学习的印刷网点补偿方法,以实现对局部网点的精准补偿,提高印刷质量和色彩还原效果。方法 通过获取待印印张的材料参数、设备参数,以及1个月内产生的具有相同参数的数字印张,利用PDC模型进行训练,生成网点补偿值,从而实现对局部网点的精准补偿。结果 实验结果表明,该方法可使网点变形率的算术平均值降低30.1%,色彩还原误差（ΔE）的算术平均值减少28.2%。此外,该方法在不同材料和设备条件下均能稳定运行,显示出良好的通用性和适应性。实验数据进一步证明了PDC模型在局部网点补偿中的精准性和高效性,为印刷行业的数字化和智能化发展提供了有力的技术支持。结论 基于机器学习的印刷网点补偿方法克服了传统补偿技术的局限性,为印刷行业的数字化和智能化发展提供了新的思路,具有广阔的应用前景。
Abstract
To address the issues of inaccurate color reproduction and declining print quality caused by dot deformation in offset printing, the work aims to propose a machine learning-based dot compensation method to achieve precise local dot compensation, thereby improving print quality and color reproduction. By acquiring material parameters and equipment parameters of the target print sheet, combined with recent digital print data within one month, the PDC model was trained to generate dot compensation values, enabling accurate local dot compensation. Experimental validation demonstrated that the proposed method reduced the arithmetic mean of dot deformation rate by 30.1% and decreased the arithmetic mean of color reproduction error (ΔE) by 28.2%. Additionally, the method exhibited stable performance across different materials and equipment conditions, demonstrating strong generalizability and adaptability. Further experimental data confirmed the precision and efficiency of the PDC model in local dot compensation, providing robust technical support for the digital and intelligent advancement of the printing industry. The machine learning-based dot compensation method overcomes the limitations of traditional compensation techniques, offering a novel approach for the digital and intelligent development of the printing industry, with broad application prospects.
关键词
印刷网点补偿 /
机器学习 /
PDC模型 /
网点变形 /
色彩还原 /
智能化印刷
Key words
print dot compensation /
machine learning /
PDC model /
dot deformation /
color reproduction /
intelligent printing
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] MA X P, CAO S Z, LI Y.Data Analysis of Print Quality Defect Detection Based on FP-Growth Algorithm[C]// 2023 5th International Conference on Frontiers Technology of Information and Computer (ICFTIC). Qingdao: IEEE, 2023: 97-100.
[2] 周喜焕. 通过网点增大补偿曲线提高印刷质量的研究[J]. 造纸装备及材料, 2023, 52(3): 1-3.
ZHOU X H.Study on Improving Printing Quality by Increasing Monson of Dots[J]. Papermaking Equipment & Materials, 2023, 52(3): 1-3.
[3] 刘琳琳, 周雨文, 王东琦, 等. 数字孪生在印刷业智能化建设中的探索与实践[J]. 印刷与数字媒体技术研究, 2024(1): 1-12.
LIU L L, ZHOU Y W, WANG D Q, et al.Exploration and Practice of Digital Twin in the Intelligent Construction of Printing Industry[J]. Printing and Digital Media Technology Study, 2024(1): 1-12.
[4] 杜宣. 浅谈工业4.0下如何实现印刷工业的智能化[J]. 印刷杂志, 2022(3): 51-54.
DU X.How to Realize the Intelligentization of Printing Industry under Industry 4.0[J]. Printing Field, 2022(3): 51-54.
[5] ZHANG K X, LIU H H, LIU X X.Research on Optical Simulation System for Halftone Printing Dot with Monte Carlo Method[J]. Optik, 2018, 156: 184-190.
[6] 胡锡金. 胶印过程中印迹变形问题的探讨[J]. 印刷技术, 2024(5): 27-30.
HU X J.Discussion on Imprint Deformation in Offset Printing Process[J]. Printing Technology, 2024(5): 27-30.
[7] 亢国军, 何永志, 王华明, 等. 基于印刷生产逻辑的订单信息流转新模式[J]. 印刷与数字媒体技术研究, 2023(6): 59-66.
KANG G J, HE Y Z, WANG H M, et al.New Model of Order Information Transfer Focusing on Printing Production Logic[J]. Printing and Digital Media Technology Study, 2023(6): 59-66.
[8] 王愿. 基于数字图像处理技术的全画面印品质量检测[D]. 西安: 西安理工大学, 2019: 33-45.
WANG Y.Quality Inspection of Full-Screen Prints Based on Digital Image Processing Technology[D]. Xi'an: Xi'an University of Technology, 2019: 33-45.
[9] 王永峰, 邱英祥, 李培杰. 浅析印刷网点变形及其控制方法[J]. 广东印刷, 2022(3): 29-30.
WANG Y F, QIU Y X, LI P J.Analysis of Printing Dot Deformation and Its Control Methods[J]. Guangdong Printing, 2022(3): 29-30.
[10] 贾志超. 边界积分方程中非结构网格划分及自适应单元细分方法研究[D]. 淄博: 山东理工大学, 2024: 10-29.
JIA Z C.Research on Unstructured Grid Division and Adaptive Element Subdivision Method in Boundary Integral Equation[D]. Zibo: Shandong University of Technology, 2024: 10-29.
[11] LI C L, YAO Z W, WU X X, et al.DeepSpeed Data Efficiency: Improving Deep Learning Model Quality and Training Efficiency via Efficient Data Sampling and Routing[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2024, 38(16): 18490-18498.
[12] 赖正良. 基于深度卷积神经网络的图像隐写分析和被隐藏信息重构研究[D]. 南昌: 南昌大学, 2023: 25-36.
LAI Z L.Research on Image Steganography Analysis and Reconstruction of Hidden Information Based on Deep Convolution Neural Network[D]. Nanchang: Nanchang University, 2023: 25-36.
[13] JEONG T, KORATIKERE P, LEIFSSON L, et al.Adaptive Hyperparameter Tuning within Neural Network-Based Efficient Global Optimization[C]// Computational Science - ICCS 2024. Cham: Springer Nature Switzerland, 2024: 74-89.
[14] FANG Z, XU X X, LI X, et al. SPGD Algorithm Optimization Based on Adam Optimizer[C]// AOPC2020: Optical Sensing and Imaging Technology. Beijing: SPIE, 2020: 115672S.1-115672S.8.
[15] GAO X Y, QIU M K, ZHAO H.Architecture Search for Deep Neural Network[C]// Smart Computing and Communication: 7th International Conference on Smart Computing and Communication (SmartCom 2022), 2023: 81-596.
[16] XI S K, ZHANG Y X.Neugebauer Reflectance Model of Frequency Modulation Halftone Image[J]. Optik, 2013, 124(15): 2103-2105.
[17] SEKINE M, SHIMBARA D, MYOJIN T, et al.Visualization Tool for Extraction of Various Attributes and Corresponding Data for Dataset Quality Assessment[C]// 2023 IEEE International Conference on Artificial Intelligence Testing (AITest). Athens: IEEE, 2023: 61-65.
[18] CYNTHIA W.On the Circuit[J]. Sound & video contractor, 2023, 41(1): 6.
PDF(426 KB)
渝公网安备 50010702501716号 渝ICP备15012534号-2
