作者：刘桂雄¹, 刘思洋¹, 吴俊芳², 罗文佳³

作者单位：1. 华南理工大学机械与汽车工程学院, 广东 广州 510640;
2. 华南理工大学物理与光电学院, 广东 广州 510640;
3. 广州市银科电子有限公司, 广东 广州 510663

关键词：机器视觉;目标检测;深度学习;卷积神经网络;票据检测

摘要：

基于深度学习的目标检测是机器视觉应用的重要方面。该文系统总结基于区域候选的目标检测算法、基于回归方法的目标检测算法及其他优化算法的算法思想、网络架构、演进过程、技术指标、应用场景，指出在机器视觉系统应用中，应充分考虑检测对象、检测精度、实时性能要求，结合不同目标检测算法特点，选择最合适的检测算法。最后，面向票据检测需求，分析目标检测算法在票据图像位置检测、防伪特征检测、文本信息检测中的应用。

Machine vision object detection algorithm based on deep learning and application in banknote detection
LIU Guixiong¹, LIU Siyang¹, WU Junfang², LUO Wenjia³
1. School of Mechanical and Automotive Engineering, South China University of Technology, Guangzhou 510640, China;
2. School of Physics, South China University of Technology, Guangzhou 510640, China;
3. Guangzhou Yin Ke Electronics Co., Ltd., Guangzhou 510663, China
Abstract: Object detection based on deep learning is an important aspect of machine vision applications. This paper systematically summarizes the object detection algorithm based on region proposals, the object detection algorithm based on regression method and the other optimization algorithm, then analyze theirs' network architecture, evolution process, technical indicators and application scenarios. It is pointed out that in the application of machine vision system, the object, detection accuracy and real-time performance requirements should be fully considered. Combining the characteristics of different object detection algorithms, the most suitable detection algorithm should be selected. Finally, for the banknote detection requirements, the application of the target detection algorithm in ticket image position detection, anti-counterfeiting feature detection and text information detection is analyzed.
Keywords: machine vision;object detection;deep learning;convolutional neural networks;banknote detection
2019, 45(5):1-9  收稿日期: 2019-03-29;收到修改稿日期: 2019-04-15
基金项目: 广州市产学研重大项目（201802030006）；广东省现代几何与力学计量技术重点实验室开放课题（SCMKF201801）
作者简介: 刘桂雄(1968-),男,广东揭阳市人,教授,博导,主要从事测控技术及仪器研究
参考文献
[1] 尹仕斌, 任永杰, 刘涛, 等. 机器视觉技术在现代汽车制造中的应用综述[J]. 光学学报, 2018, 38(8):11-22
[2] GUO Y, LIU Y, OERLEMANS A, et al. Deep learning for visual understanding:A review[J]. Neurocomputing, 2016, 187:27-48
[3] FELZENSZWALB P F, GIRSHICK R B, MCALLESTER D, et al. Object detection with discriminatively trained part-based models[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2010, 32(9):1627-1645
[4] REN X, RAMANAN D. Histograms of sparse codes for object detection[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2013.
[5] FELZENSZWALB P, GIRSHICK R, MCALLESTER D, et al. Visual object detection with deformable part models[J]. Communications of the ACM, 2013, 56(9):97-105
[6] DALAL N, TRIGGS B. Histograms of oriented gradients for human detection[C]//International Conference on Computer Vision & Pattern Recognition (CVPR'05), 2005.
[7] LOWE D G. Distinctive image features from scale-invariant keypoints[J]. International Journal of Computer Vision, 2004, 60(2):91-110
[8] CHEN P, LIN C, SCHÖLKOPF, B. A tutorial on v-support vector machines[J]. Applied Stochastic Models in Business & Industry, 2005, 21(2):111-136
[9] WANG P, SHEN C, BARNES N, et al. Fast and robust object detection using asymmetric totally corrective boosting[J]. IEEE Transactions on Neural Networks and Learning Systems, 2012, 23(1):33-46
[10] 尹宏鹏, 陈波, 柴毅, 等. 基于视觉的目标检测与跟踪综述[J]. 自动化学报, 2016, 42(10):1466-1489
[11] 方路平, 何杭江, 周国民. 目标检测算法研究综述[J]. 计算机工程与应用, 2018, 54(13):11-18,33
[12] 于进勇, 丁鹏程, 王超. 卷积神经网络在目标检测中的应用综述[J]. 计算机科学, 2018, 45(S2):17-26
[13] 张明江, 李红卫, 赵卫虎, 等. 深度学习在军用光缆线路无人机巡检中的应用[J]. 光通信研究, 2018(6):61-65
[14] 常海涛, 苟军年, 李晓梅. Faster R-CNN在工业CT图像缺陷检测中的应用[J]. 中国图象图形学报, 2018, 23(7):129-139
[15] 王卫东, 程丹. 监控场景下的实时车辆检测方法[J]. 电子测量与仪器学报, 2018(7):83-88
[16] 魏震宇, 文畅, 谢凯, 等. 光流估计下的移动端实时人脸检测[J]. 计算机应用, 2018, 38(4):1146-1150
[17] 张玉杰, 张媛媛. 便携式票据数字水印检测系统的研究[J]. 自动化仪表, 2013, 34(3):41-43
[18] 陶锐, 孙彦景. 金融票据混沌水印加密算法研究与实现[J]. 电子器件, 2017, 40(5):1297-1303
[19] PHAM T D, NGUYEN D T, KIM W, et al. Deep learning-based banknote fitness classification using the reflection images by a visible-light one-dimensional line image sensor[J]. Sensors, 2018, 18(2):472
[20] LEE J, HONG H, KIM K, et al. A survey on banknote recognition methods by various sensors[J]. Sensors, 2017, 17(2):313
[21] PHAM T, LEE D, PARK K. Multi-national banknote classification based on visible-light line sensor and convolutional neural network[J]. Sensors, 2017, 17(7):1595
[22] LECUN Y, BENGIO Y, HINTON G. Deep learning[J]. Nature, 2015, 521(7553):436
[23] EVERINGHAM M, ESLAMI S M A, VAN GOOL L, et al. The pASCAL visual object classes challenge:a retrospective[J]. International Journal of Computer Vision, 2015, 111(1):98-136
[24] 陈超, 齐峰. 卷积神经网络的发展及其在计算机视觉领域中的应用综述[J]. 计算机科学, 2019, 46(3):63-73
[25] KRIZHEVSKY A, SUTSKEVER I, HINTON G E. Imagenet classification with deep convolutional neural networks[C]//Advances in Neural Information Processing Systems,2012.
[26] SZEGEDY C, LIU W, JIA Y, et al. Going deeper with convolutions[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition,2015.
[27] SIMONYAN K, ZISSERMAN A. Very deep convolutional networks for large-scale image recognition[C]//2015 International Conference on Learning Representations, 2015.
[28] HE K, ZHANG X, REN S, et al. Deep residual learning for image recognition[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2016.
[29] 张荣, 李伟平, 莫同. 深度学习研究综述[J]. 信息与控制, 2018, 47(4):385-397,410
[30] 吴帅, 徐勇, 赵东宁. 基于深度卷积网络的目标检测综述[J]. 模式识别与人工智能, 2018, 31(4):335-346
[31] 吴加莹, 杨赛, 堵俊, 等. 自底向上的显著性目标检测研究综述[J]. 计算机科学, 2019, 46(3):48-52
[32] ZHAO Z Q, ZHENG P, XU S, et al. Object detection with deep learning:A review[C]//IEEE Transactions on Neural Networks and Learning Systems, 2019.
[33] GIRSHICK R, DONAHUE J, DARRELL T, et al. Rich feature hierarchies for accurate object detection and semantic segmentation[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2014.
[34] HE K, ZHANG X, REN S, et al. Spatial pyramid pooling in deep convolutional networks for visual recognition[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2015, 37(9):1904-1916
[35] GIRSHICK R. Fast R-CNN[C]//Proceedings of the IEEE International Conference on Computer Vision,2015.
[36] RUMELHART D E. Learning representations by back-propagating errors[J]. Nature, 1986, 323(6088):533-536
[37] REN S, HE K, GIRSHICK R, et al. Faster R-CNN:Towards real-time object detection with region proposal networks[C]//Advances in Neural Information Processing Systems,2015.
[38] REDMON J, DIVVALA S, GIRSHICK R, et al. You only look once:Unified, real-time object detection[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition,2016.
[39] REDMON J, FARHADI A. YOLO9000:better, faster, stronger[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition,2017.
[40] IOFFE S, SZEGEDY C. Batch normalization:accelerating deep network training by reducing internal covariate shift[C]//International Conference on Machine Learning, 2015.
[41] REDMON J, FARHADI A. Yolov3:An incremental improvement[J].CoRR, 2018.abs/1804.02767.
[42] SZEGEDY C, VANHOUCKE V, IOFFE S, et al. Rethinking the inception architecture for computer vision[C]//Proceeding of the IEEE Conference on Computer Vision and Pattern Recognition,2016.
[43] LIU W, ANGUELOV D, ERHAN D, et al. Ssd:Single shot multibox detector[C]//European Conference on Computer Vision, 2016.
[44] LIN T Y, GOYAL P, GIRSHICK R, et al. Focal loss for dense object detection[C]//Proceedings of the IEEE International Conference on Computer Vision, 2017.
[45] ZHANG S, WEN L, BIAN X, et al. Single-shot refinement neural network for object detection[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018.
[46] 葛动元, 姚锡凡, 向文江, 等. 面向齿廓偏差等精密检测的机器视觉关键技术[J]. 机械传动, 2019, 43(2):171-176
[47] 郭雪梅, 刘桂雄, 黄坚, 等. 面向标准件装配质量的PI-SURF检测区域划分技术[J]. 中国测试, 2017, 43(8):101-105
[48] 广州市银科电子有限公司. 一种基于红外油墨标志智能识别的票据防伪鉴别方法:CN201710536627.2[P]. 2017-11-10.
[49] YOUNG P, SEUNG K, TUYEN P, et al. A high performance banknote recognition system based on a one-dimensional visible light line sensor[J]. Sensors, 2015, 15(6):14093-14115
[50] LIU X W, LIU C Y. Paper currency CIS image fuzzy enhancement and boundary detection[J]. Applied Mechanics and Materials, 2014, 651-653:2356-2361
[51] 广州市银科电子有限公司. 基于防伪材料光谱特性的票据鉴伪方法及装置:CN201710516001.5[P]. 2017-10-24.
[52] 广州市银科电子有限公司. 一种智能识别水印特征的票据防伪鉴别方法:CN201710337615.7[P]. 2017-09-05.
[53] ROY A, HALDER B, GARAIN U, et al. Machine-assisted authentication of paper currency:an experiment on Indian banknotes[J]. International Journal on Document Analysis and Recognition (IJDAR), 2015, 18(3):271-285
[54] BRUNA A, FARINELLA G, GUARNERA G, et al. Forgery detection and value identification of Euro banknotes[J]. Sensors, 2013, 13(2):2515-2529
[55] LIAO M, SHI B, BAI X, et al. Textboxes:A fast text detector with a single deep neural network[C]//Proceedings of the Thirty-First AAAI Conference on Artificial Intelligence,2017.