随着国民生活水平提升，消费者对包装印刷的品质提出了更高要求。目前传统光电与超声波类传感器纠偏系统检测精度较低，视觉类纠偏系统的额外印刷标记会造成材料浪费。通过分析现有纠偏系统的问题，本文提出一种以视觉为反馈并利用印刷图案来计算偏移量的纠偏系统方案。

（1）从包装印刷品特点出发，确立了目标图案的选取规则。考虑系统实时性的要求，选择计算简单、抗噪性强的傅里叶描述子用于目标图案识别。针对傅里叶描述子对起始点敏感导致识别准确率低的问题，在获取到图像的轮廓信息后，利用二阶中心矩的方向特性计算过轮廓形心的轴线，将轴线与轮廓的交点作为傅里叶描述子的起始点。实验证明改进后识别准确率达到84%，相较改进前提高了16%。

（2）针对识别到目标后像素级定位精度低的问题，采用Zernike矩亚像素边缘检测算法提高目标定位精度。为了解决传统Zernike矩检测算法边缘判定阈值选取不当导致边缘丢失和伪边缘的问题，通过最大类间方差法计算亚像素边缘最佳阶跃阈值，简化传统方法手动调节阈值的繁琐步骤，同时将目标定位精度提升至亚像素级。

（3）通过最小二乘拟合法自动拟合基准线，当发生偏移时通过目标图案到基准线的距离计算偏移量，以此作为视觉反馈，利用模糊控制实现对PID参数的在线动态调整，经过实验验证，此方法对偏移修正具有较好的控制效果。

结合相关算法的研究搭建纠偏系统模型，对系统的精度和实时性两方面进行测试。结果表明，纠偏最大误差在0.03 mm内，平均纠偏误差为0.0127 mm，符合国家标准。在本文选定的检测区域内，系统的检测帧率为平均每秒50.15帧，具有较好的实时性。本文提出的纠偏方法有较高的纠偏精度和实时性，具有工程应用价值。

With the improvement of national living standard, consumers have put forward higher requirements on the quality of packaging printing. At present, the detection accuracy of traditional photoelectric and ultrasonic sensor-based offset correction system is low, and the extra printing marks of vision-based offset correction system will cause material waste. By analyzing the problems of existing guiding systems, this paper proposes a guiding system solution that uses vision as feedback and uses printing patterns to calculate the offset amount.

(1) The selection rules of the target pattern are established from the characteristics of packaging printed materials. Considering the requirement of system real-time, the Fourier descriptor with simple calculation and strong noise immunity is selected for target pattern recognition. For the problem that Fourier descriptor is sensitive to the starting point, the recognition accuracy is low. After obtaining the contour information of the image, the directional property of the second-order central moment is used to calculate the axis over the center of the contour, and the intersection of the axis and the contour is used as the starting point of the Fourier descriptor. It is demonstrated that the recognition accuracy reaches 84% after the improvement, which is 16% higher than that before the improvement.

(2) To address the problem of low pixel-level localization accuracy after recognizing the target, the Zernike moment sub-pixel edge detection algorithm is used to improve the target localization accuracy. In order to solve the problem of edge loss and pseudo-edge caused by the improper selection of the edge determination threshold of the traditional Zernike moment detection algorithm, the optimal step threshold of subpixel edge is calculated by the maximum inter-class variance method, which simplifies the tedious steps of manually adjusting the threshold by the traditional method and improves the target localization accuracy to the subpixel level at the same time.

(3) The baseline is automatically fitted by the least squares fitting method, and the offset is calculated by the distance from the target pattern to the baseline when the offset occurs, which is used as the visual feedback to realize the online dynamic adjustment of PID parameters by using fuzzy control.

Combined with the study of related algorithms to build a model of the deflection correction system, the system is tested in terms of both accuracy and real-time performance. The results show that the maximum error of deflection correction is within 0.03 mm, and the average error of deflection correction is 0.0127 mm, which is in line with the national standard. In the selected detection area of this paper, the detection frame rate of the system is 50.15 frames per second on average, which has good real-time performance. The correction method proposed in this paper has high correction accuracy and real-time performance, and has engineering application value.

[1]中商情报网. 2022年中国包装行业市场规模预测及其细分市场占比分析[EB/OL]. https://m.askci.com/news/chanye/20220616/1538351892764.shtml.

[2]中国轻工业信息网. 2022年我国纸包装行业出口规模将达79.44亿美元[EB/OL]. http://www.clii.com.cn/zhhylm/zhhylmHangYeJuJiao/202209/t20220915_3954974.html.

[3]王佳, 武淑琴, 王仪明, 等. 基于数字图像处理的套印精度检测方法研究[J]. 北京印刷学院学报, 2021, 29(06): 149-154.

[4]袁沛沛. 机电自动化中传感器技术的应用[J]. 南方农机, 2020, 51(07): 235.

[5]牛伟锋, 宗亮亮, 张小康, 等. 基于颜色传感器定位的线激光罐道结构检测精度提高方法[J]. 山西焦煤科技, 2021, 45(10): 18-22.

[6] 龙俊. 纸质包装印刷中的数字印刷技术应用探讨[J]. 造纸信息, 2021(06): 95-96.

[7]肖萌. 浅析印刷产品质量检验人员应具备的几种能力[J]. 广东印刷, 2021(01): 47-48.

[8]Kim J, Kim Y, Kim T, et al. Printing pressure uniformization through adaptive feedforward control in roll-to-roll printing process[J]. Microsystem Technologies, 2020, 2: 265-273.

[9]Kim T, Kim Y, Kim J, et al. A precise position measurement system of roll-to-roll printing using burst alignment patterns[J]. Microsystem Technologies, 2020, 26(1): 231-238.

[10]邓瑞, 侯和平, 徐卓飞, 等. 印刷纸带纠偏装置研究与系统开发[J]. 包装工程, 2018, 39(13): 138-143.

[11]Wang H, Du H, Cui Q, et al. Artificial bee colony algorithm based PID controller for steel stripe deviation control system[J]. Science progress, 2022, 105(1).

[12]Sun X, Wang Y. Research on Automatic Deviation Correction Device and Control System for Prevent Conveyor Belt Deviation in the Curve Section[J]. Journal of Failure Analysis and Prevention, 2022, 22(6): 2272-2287.

[13]Xiao Y, Zhao C, Qi H, et al. Research on the optimization strategy of lithium battery electrode correction system based on algorithm fusion[J]. Journal of Intelligent & Fuzzy Systems, 2023, 44(2): 2503-2523.

[14]胡亚南, 霍蛟飞, 赵慧娟, 等. 食品包装袋膜模糊单神经元PID纠偏控制设计[J]. 制造业自动化, 2022, 44(05): 56-60.

[15]曾飞, 宋杰杰, 胡文祥, 等. 带式输送机输送带横向纠偏控制系统[J]. 现代制造工程, 2021, 492(09): 107-111+117.

[16]Xu C, Yang X, He Z, et al. Precise Positioning of Circular Mark Points and Transistor Components In SMT Applications[J]. IEEE Transactions on Industrial Informatics, 2020, 6: 1-11.

[17]刘斌, 董正天, 胡春海, 等. 基于机器视觉的丝网印刷样板尺寸测量方法[J]. 计量学报, 2021, 42(2): 150-156.

[18]巨茗升, 王伟乾. 贴片机视觉系统设计与分析[J]. 轻工科技: 2019, 35(7): 72-74.

[19]贺香华, 陈从桂, 周聪, 等. 基于机器视觉的点胶质量检测方法[J]. 组合机床与自动化加工技术, 2020, 561(11): 99-101+106.

[20]Panetta K, Kamath S, Rajeev S, et al. LQM: Localized Quality Measure for Fingerprint Image Enhancement[J]. IEEE Access, 2019, 7: 104567-104576.

[21]Li C, Zhang X, Huang Y, et al. A Novel Algorithm for Defect Extraction and Classification of Mobile Phone Screen Based on Machine Vision[J]. Computers & Industrial Engineering, 2020, 146.

[22]Sinha A, Garg N, Singh N, et al. Space robotics hybrid conceptual model for tracking and estimation using IoRT and AI-based control system[J]. Materials Today: Proceedings, 2023, 72(3): 958-965.

[23]陈耀欢, 任德均, 邓霖杰, 等. 基于机器视觉的纠偏检测系统[J]. 机电工程技术, 2014, 43(11): 86-88.

[24]李婧. 凹版印刷多色套准控制系统设计[J]. 包装工程, 2021, 42(19): 277-281.

[25]裴茂印, 张荣福, 曾忠, 等. 柔版印刷套印误差检测技术研究[J]. 包装工程, 2019, 40(23): 245-250.

[26]李硕, 杜媛. 基于嵌入式技术的印刷设备套色精定位方法研究[J]. 自动化与仪器仪表, 2020(06): 128-131.

[27]陈锦华, 何邦贵, 杨述鑫, 等. 镭射光柱纸凹印套准色标位置优化及实现[J]. 包装工程, 2021, 42(21): 188-193.

[28]吕明珠, 吴学毅, 成刚虎, 等. 基于K-means聚类十字线的套印偏差检测方法研究[J]. 包装工程, 2020, 41(01): 143-148.

[29]Wang Y, Qi Y, Zhang Y, et al. Optimization of overprint error detection based on the FCM algorithm[C]//2022 International Conference on Image Processing, Computer Vision and Machine Learning. IEEE, 2022: 611-614.

[30]吕俊霞, 焦欣欣. 鲜切蔬菜包装过程袋膜纠偏控制系统设计[J]. 包装工程, 2019, 40(21): 205-210.

[31]赵璐, 杨伟兵. 小袋食品包装机袋膜纠偏控制系统设计[J]. 食品工业, 2020, 41(02): 210-213.

[32]申九菊. 枕式包装机纠偏控制系统设计[J]. 包装工程, 2022, 43(13): 231-237.

[33]黄阮莹, 晁季蕾 ,褚夫强, 等. 包装印刷技术应用方案及发展趋势[J]. 丝网印刷, 2021, 320(12): 59-63.

[34]李天楠. 皮带机跑偏原因及防偏技术探讨[J]. 当代化工研究, 2022, 22(21): 132-134.

[35]Cheng F. Application Analysis of Digital Printing Technology in Packaging Printing[J]. E3S Web of Conferences, 2020, 185: 02034-02037.

[36]Gurjinder S, Amandeep K. Artificial Bee Colony Optimized Multi-Histogram Equalization for Contrast Enhancement and Brightness Preservation of Color Images[J]International Journal of Engineering and Manufacturing, 2023, 13(1): 45-58.

[37]黄梦涛, 连一鑫.基于改进Canny算子的锂电池极片表面缺陷检测[J]. 仪器仪表学报, 2021, 42(10): 199-209.

[38]Ma S, Yang C, Bao S. Contrast Enhancement Method Based on Multi-Scale Retinex and Adaptive Gamma Correction[J]. Journal of Advanced Computational Intelligence and Intelligent Informatics, 2022, 26(6): 875-883.

[39]胡钰, 李甜甜, 黄梁松, 等. 基于双边伽马校正的保亮度图像增强方法[J]. 计算机应用与软件, 2019, 36(05): 204-210+241.

[40]焦瑾瑾. 基于Roberts边缘检测的数字媒体激光重现方法[J]. 激光杂志, 2022, 43(05): 197-200.

[41]Saeed B, Hossein B. Edge detection on noisy images using Prewitt operator and fractional order differentiation[J]. Multimedia Tools and Applications, 2022, 81(7): 9759-9770.

[42]Chang Q, Li X, Li Y, et al. Multi-directional Sobel operator kernel on GPUs[J]. Journal of Parallel and Distributed Computing, 2023, 177: 160-170.

[43]魏雨, 黄玉蕾. Kirsch联合高低双阈值的RGB图像边缘检测算法[J]. 计算机测量与控制, 2023, 31(03): 95-101.

[44]Naji O, Shah H, Anwar N, Johan N. Square groove detection based on Förstner with Canny edge operator using laser vision sensor[J]. The International Journal of Advanced Manufacturing Technology, 2023, 125(5-6): 2885-2894.

[45]吴琼, 马雷. 一种基于LoG算子的量子图像边缘检测算法[J]. 量子电子学报, 2022, 39(05): 720-727.

[46]Amal H, Jaouad E, Mostafa J. Rotation scaling and translation invariants by a remediation of Hu’s invariant moments[J]. Multimedia Tools and Applications, 2020, 79: 1-39.

[47]白鑫, 卫琳. 基于双级特征提取与度量的图像检索算法[J]. 包装工程, 2018, 39(21): 198-205.

[48]王福斌, 潘兴辰, 孙志林. 飞秒激光烧蚀光斑轮廓提取及其链码描述研究[J]. 中国测试, 2020, 46(04): 116-122.

[49]李云红, 罗雪敏, 苏雪平, 等. 基于改进曲率尺度空间算法的电力设备红外与可见光图像配准[J]. 激光与光电子学进展, 2022, 59(12): 138-145.

[50]Tigistu T, Abebe G. Classification of rose flowers based on Fourier descriptors and color moments[J]. Multimedia Tools and Applications, 2021, 80(30): 1-15.

[51]赵莉, 张克旺. 基于傅里叶描述子的实装战技图像典型特征相似度检测方法[J]. 南京理工大学学报, 2022, 46(04): 406-411.

[52]高泽. 基于改进傅里叶描述子的商品图像检索技术研究[D]. 哈尔滨: 哈尔滨商业大学, 2019.

[53]吴一全, 邹宇, 刘忠林. 基于Franklin矩的亚像素级图像边缘检测算法[J]. 仪器仪表学报, 2019, 40(05): 221-229.

[54]Mo J, Yan H, Liu J. An adaptive sub-pixel edge detection method based on improved Zernike moment[J]. International Journal of Wireless and Mobile Computing, 2022, 22(2): 140-147.

[55]艾双哲, 段发阶, 李杰, 等. 形心匹配优化下的狭长空间近距离轨迹测量系统[J]. 红外与激光工程, 2023, 52(03): 259-271.

[56]许洋, 周奎, 杨亚会, 等. 基于增量式PID的步进电机速度控制系统设计[J]. 科技与创新, 2022, 193(01): 172-175+178.

[57]段兴尧. 基于模糊控制实现采煤机的自动化调速控制[J]. 机械管理开发, 2022, 37(10): 100-102.

[58]高歆杨, 柯芳, 邹伟, 等. 基于模糊控制策略的快速反射镜伺服控制[J]. 兵工学报, 2020, 41(08): 1529-1538.

[59]Shen C, Ma C, Gao W. Design of a Ultra-Stable Low-Noise Space Camera Based on a Large Target CMOS Detector and Image Data Analysis[J]. Sensors, 2022, 22(24): 9991.

[60]张占军, 张磊, 米士隆. 数码相机镜头的使用与维护[J]. 中国设备工程, 2020, 454(18): 59-60.

[61]史小星. 基于热仿真的工业用光源寿命设计与分析[J]. 机电产品开发与创新, 2021, 34(04): 23-24+53.

[62]曾志强, 冯鹏鹏, 李忠华. 影像测量中大视场相机畸变精确校正研究[J]. 机床与液压, 2022, 50(02): 21-25.

[63]王谭, 王磊磊, 张卫国, 等. 基于张正友标定法的红外靶标系统[J]. 光学精密工程, 2019, 27(08): 1828-1835.

[64]曾骏, 何剑锋, 李卫东, 等. 基于ARM-Linux/Qt便携式智能γ能谱仪软硬件设计与实现[J]. 仪表技术与传感器, 2022, 478(11): 17-22+28.

[65]徐鑫, 李春晖, 赵丽清, 等. 基于OpenCV的施肥铲精准对行系统设计[J]. 中国农机化学报, 2022, 43(12): 142-147.