纸币清分机图像识别与残缺检测方法研究
|
摘要
目前在世界上每天流通的纸币数量巨大,在金融部门内部纸币整理工作是非常繁重的,如何快速准确的纸币清分在银行业中具有非常重要的意义,通过使用灵敏准确的纸币清分机,能使繁琐的钞票清分工作变得简易、快捷和可靠。
清分系统的核心技术基础是实时纸币图像处理和图像识别。清分软件对输入的纸币图像进行处理,计算出清分结果并送给清分控制装置,由清分控制装置根据清分结果和清分机当前的运行状态完成清分动作。在清分机的系统中,对实时性要求非常高,也就是纸币经过图像传感器的时候,必须在一定的时间内计算出纸币的面额、面向、朝向等信息,我们选用的距离分类器,它的特性是计算简单,速度快,这也正是我们需要的。
本文主要是针对各种外币的特点,在原来人民币清分系统的基础上,作了几点改进:
(1)根据多币种的不同特性和时间的要求,给出了快速与抗干扰特征向量的提取的新方法,与原系统和其他成分数的选择方法相比,不但速度,而且具有更高的识别率。
(2)外币的种类的多,为了能够进行多币种的分类,根据不同币种我们给出了多级分类器的新方法,该方法根据不同币种,自动选择与组合分类器,最终达到准确清分多种外币的目的。
(3)针对纸币的残缺清分需求,给出了一种基于图像边缘特征的污损检测方法,力图有效地检测出纸币上的笔划及撕裂等污损特征,并计算出相应的污损面积。
清分程序与控制程序均用C语言编写,高效快速,便于维护和升级。将文中实现的纸币清分机软件系统在浙江金利电子有限公司的JVC-200机型上实际应用表明,该软件系统可以很好地满足需求。
There are large numbers of paper currency circulates in the worldwide every day, it is a tough work for staves to sort the paper currency in the financial condition. How to sort the paper currency quickly and correctly becomes very important in the banks. The technology is to process and recognize national paper currency.
The key technology of the sorter is real-time image processing. Classifier processes the note image, and then sends the result of the paper currency to controlling system. The machine takes corresponding action to finish the classifying according to the answer. In order to speed up our machine to sort the currency quickly, we select shortest-distance classifier whose feature is computation easy and speed high.
We make several improvements according to the differences of different foreign paper currency.
(1) Feature extraction is improved. Compared with original method, we give a fast and anti-interference feature extraction. The advantage of this method is that it has a faster speed and better classifying result.
(2) There are many kinds of foreign paper currency. In order to classify different kinds of foreign paper currency, we give a multistage classifier. This new method will select and combo different classifiers based on different foreign paper currency and classifies paper currency correctly.
(3) Defect detection improvement .An edge-based algorithm is proposed to detect the scratches and cracks appearing frequently on paper currency.
The system of the control and classy is programmed with c language. The benefit is quick developing and easy upgrading. We apply the method in this article to the JVC-200 machine and receive a good result.
清分系统的核心技术基础是实时纸币图像处理和图像识别。清分软件对输入的纸币图像进行处理,计算出清分结果并送给清分控制装置,由清分控制装置根据清分结果和清分机当前的运行状态完成清分动作。在清分机的系统中,对实时性要求非常高,也就是纸币经过图像传感器的时候,必须在一定的时间内计算出纸币的面额、面向、朝向等信息,我们选用的距离分类器,它的特性是计算简单,速度快,这也正是我们需要的。
本文主要是针对各种外币的特点,在原来人民币清分系统的基础上,作了几点改进:
(1)根据多币种的不同特性和时间的要求,给出了快速与抗干扰特征向量的提取的新方法,与原系统和其他成分数的选择方法相比,不但速度,而且具有更高的识别率。
(2)外币的种类的多,为了能够进行多币种的分类,根据不同币种我们给出了多级分类器的新方法,该方法根据不同币种,自动选择与组合分类器,最终达到准确清分多种外币的目的。
(3)针对纸币的残缺清分需求,给出了一种基于图像边缘特征的污损检测方法,力图有效地检测出纸币上的笔划及撕裂等污损特征,并计算出相应的污损面积。
清分程序与控制程序均用C语言编写,高效快速,便于维护和升级。将文中实现的纸币清分机软件系统在浙江金利电子有限公司的JVC-200机型上实际应用表明,该软件系统可以很好地满足需求。
There are large numbers of paper currency circulates in the worldwide every day, it is a tough work for staves to sort the paper currency in the financial condition. How to sort the paper currency quickly and correctly becomes very important in the banks. The technology is to process and recognize national paper currency.
The key technology of the sorter is real-time image processing. Classifier processes the note image, and then sends the result of the paper currency to controlling system. The machine takes corresponding action to finish the classifying according to the answer. In order to speed up our machine to sort the currency quickly, we select shortest-distance classifier whose feature is computation easy and speed high.
We make several improvements according to the differences of different foreign paper currency.
(1) Feature extraction is improved. Compared with original method, we give a fast and anti-interference feature extraction. The advantage of this method is that it has a faster speed and better classifying result.
(2) There are many kinds of foreign paper currency. In order to classify different kinds of foreign paper currency, we give a multistage classifier. This new method will select and combo different classifiers based on different foreign paper currency and classifies paper currency correctly.
(3) Defect detection improvement .An edge-based algorithm is proposed to detect the scratches and cracks appearing frequently on paper currency.
The system of the control and classy is programmed with c language. The benefit is quick developing and easy upgrading. We apply the method in this article to the JVC-200 machine and receive a good result.
引文
1 F. Takeda, S. Omatu, S. Onami, T. Kadono, and K. Terada. A Paper CurrencyRecognition Method by A Small Size Neural Network with Optimized Masks by GA. In World Congress on Computational Intelligence. 1994, 9(7):4243~4246
2 M. Tanaka. Modeling of Mixtures of Principal Component Analysis Model with Genetic Algorithm Proceeding 31st International Symposium on Stochastic Systems Theory and Its Application. 1999,6(8): 157~162
3 F. Takeda, S. Omatu and S. Onami, Mask Optimization by Genetic Algorithm for A Neuro-Pattern Recognition Machine with Masks. Trans. Inst. of Syst. Contr. and Inform. Engineers. 1999, 8(5): 196~203,
4吴浩扬等.遗传算法的一种特例-正交试验设计法.软件学报. 2001, 12(1): 148-153
5漆远等.基于遗传算法的指纹图匹配及应用.软件学报. 2000, 11(4):488~493
6杨启文 等. 遗传算法优化速度的改进. 软件学报. 2001, 12(2)
7 Xi,Yu-geng, Cai Tian-you, Yun Wei-min. Survey on genetic algorithm. Control Theory and Application. 1996, 13(6): 696~704
8 Dempsteer A P, Laird N M, Rubin D B. Maximum likelihood from incomplete data via the EM algorithm. Journalof Royal Statistical Society, Series B. 1977, 39(1):1~38
9戴葵. 神经网络实现技术. 国防科技大学出版社. 1998
10郭娟,林东,戚文雅.基于加权Fisher准则的先行鉴别分析与人脸识别
11孙即祥.模式识别中的特征提取与计算机视觉不变量.国防工业出版社, 2001年
12 A.Ahmadi, S.Omatu. A High Reliability Method for Classification of Paper Currency Based on Neural Networks Proceeding of The Eighth International Symposium on Artificial Life and Robotics. 2003, 2(8): 601~604
13 Sonehara. An application of neural network to image data. Trans. Inst.Syst, Control and Info. Eng. 1991, 35(1):11~18
14 S.Amari. The Present and Future of Neural Network. Kyoritsu Press. 1990, inJapanese
15 P. Moerland. A Comparison of Mixture Models for Density Estimation Artifical Neural Networks. Conference Publication. 1999,(470):7~10
16 F.Takeda, Sigeru Omatu. High Speed Paper Currency Recognition by Neural Networks. IEEE Trans Neural Network. 1995, l6(1)
17 Frosini, M. Gori, P. Priami. A Neural Network-Based Model for Paper Currency Recognition and Verification. IEEE Transactions on Neural Networks. 1996, 7(6):1482~1490
18李继飞.神经网络模式识别技术在自动售货机中的应用研究. 湖南大学硕士学位论文. 2001
19 F. Takeda, S. Omatu. High speed paper currency recognition by neural networks. IEEE Trans on Neural Network. 1995, 6 (1):73~77
20 F. Takeda, S. Omatu, T. Inoue, and S. Onami. A Structure Reduction of Neural Network with Random Masks and Bill Money Recognition. Proceedings of the 2nd International Conference on Fuzzy Logic and Neural Networks. 1999, 12(5):
21 H.D. Cheng, Y.Sun. A hierarchical approach to color image segmentation using homogeneity. IEEE Trans on Image Processing. 2000, 9(12):2071~2082
22 TMS320C6713B FLOATING-PINT DIGITAL SIGNAL PROCESSOR
23 TMS320C6000 ON CHIP SUPPORT LIBRARY API USER’S GUIDE
24 TMS320C6X Peripheral Support Library Programmer’s Guide
25 TMS320C6000 McBSP: UART
26 TMS320C6000 External Memory Interface Reference Guide
27 T. Michael, U.E. Martin. Digital Processor for Wavelet-based Image Compression. Proceedings of SPIE - The International Society for Optical Engineering. 2000, (3965):340~347
28徐婉莹,刘建军,黄新生.基于CPLD和DSP的高速图像采集技术研究. 电子工程师. 2004, 30(06):48~50
29 Y.L. Li, G.Q. Ni, Chen Xiao-Mei. Control for Image Fusion at Real Time Implemented with Programmable Devices. Opto-Electronic Engineering. 2003, 30(4):11~14
30 Complete 14-bit CCD/CIS Signal Processor
31杨洁,冯力刚,蒋加伏. 基于小波包和支持向量机的人脸识别. 2004, 21(9): 131~133
32 Cygnal Integrated Products,INC.潘琢金,孙德龙,夏秀峰译. C8051F单片机应用解析.北京航空航天大学出版社. 2002,10
33崔明星.纸币识别算法的研究.哈尔滨工业大学硕士学位论文. 2002
34刘家锋, 刘松波, 唐降龙.一种实时纸币识别方法的研究. 计算机研究与发展. 2003,40(7): 1057~1061
35 M. Moerland. A Comparision of Mixture Models for Density Estimation Artificial Neural Networks. Conference Publication IEE. 1999,(470):7-10
36 U.V. Chaudhari, J. Navratil. Multigrained Modeling with Pattern Specific Maximum Likelihood Transformations for Text-Independent Speaker Recognition IEEE Transactions on Speech and Audio Processing. 2003, 11(1):61~69
37 Bhattacharyya, T. Srikanthan, P. Krishnamurthy. Ideal GMM Parameters & Posteriorlog Likelihood for Speaker Verification. IEEE 2001, 3(5):208~213
38 Hedelin , J. Skoglund. Vector Quantization Based on Gaussian Mixture Models IEEE Trans. Speech and Audio Processing. 2000, 8(4):385~401
39 A. Dempster, N. Laird, D. Rubin. Maximum Likelihood from Incomplete Data via the EM Algorithm, J. Roy. Stat. Soc. 1977, 39(3):1~38
40刘关立. 任意文本的说话人识别系统研究. 大连理工大学硕士论文 2000.3
41 Dempster, A.P.Laird, N.M. and Rubin. Maximum-Lilelihood from Incomplete Data Via The EM Algorithm Journal of Royal Statistics cs Society Series. 1977, 99(1):1~38
42 J.A.Bilmes. A Gentle Tutorial of the EM Algorithm and its Application to Parameter Estimation for Gaussian Mixed and Hidden Markov Models U.C.Berkely, TR-97-021, 1998
43 Fumiaki Takeda, Sigeru Omatu. High speed paper currency recognition by neural networks. IEEE Trans on Neural Network. 1995, 6(1):73~77
44 A Frosini, M Gori, P Priami. A neural network based model for paper currency recognition and verification. IEEE Trans on Neural Networks. 1996, 7(6): 1482~1490
45 J.F. Liu, S.B. Liu, and X.L. Tang. An Algorithm of Real Time Paper Currency Recognition. Journal of computer research and development. 2003, 140(17):1057～1061
46 A. Latif-Amet, A. Ertu’zu’n, A. Erc?il. An efficient method for texture defectdetection: sub-band domain co-occurrence matrices. Image and Vision Computing . 2000, (18):543~553
47 H. Onishi, S. Tatsumi. A Pattern Defeat Inspection Method by Parallel Grayscale Image Comparison without Precise Image Alignment. IEEE 28th Annual Conference of the Industrial Electronics Society. 2002, 3(5-8): 2208~2213
48 L.G. Brown. A survey of image registration techniques. ACM Computing Surveys. 1992, 24(4):325~376
49 B. Zitová, J. Flusser. Image registration methods: a survey. Image and Vision Computing. 2003, 21(11):977~1000
2 M. Tanaka. Modeling of Mixtures of Principal Component Analysis Model with Genetic Algorithm Proceeding 31st International Symposium on Stochastic Systems Theory and Its Application. 1999,6(8): 157~162
3 F. Takeda, S. Omatu and S. Onami, Mask Optimization by Genetic Algorithm for A Neuro-Pattern Recognition Machine with Masks. Trans. Inst. of Syst. Contr. and Inform. Engineers. 1999, 8(5): 196~203,
4吴浩扬等.遗传算法的一种特例-正交试验设计法.软件学报. 2001, 12(1): 148-153
5漆远等.基于遗传算法的指纹图匹配及应用.软件学报. 2000, 11(4):488~493
6杨启文 等. 遗传算法优化速度的改进. 软件学报. 2001, 12(2)
7 Xi,Yu-geng, Cai Tian-you, Yun Wei-min. Survey on genetic algorithm. Control Theory and Application. 1996, 13(6): 696~704
8 Dempsteer A P, Laird N M, Rubin D B. Maximum likelihood from incomplete data via the EM algorithm. Journalof Royal Statistical Society, Series B. 1977, 39(1):1~38
9戴葵. 神经网络实现技术. 国防科技大学出版社. 1998
10郭娟,林东,戚文雅.基于加权Fisher准则的先行鉴别分析与人脸识别
11孙即祥.模式识别中的特征提取与计算机视觉不变量.国防工业出版社, 2001年
12 A.Ahmadi, S.Omatu. A High Reliability Method for Classification of Paper Currency Based on Neural Networks Proceeding of The Eighth International Symposium on Artificial Life and Robotics. 2003, 2(8): 601~604
13 Sonehara. An application of neural network to image data. Trans. Inst.Syst, Control and Info. Eng. 1991, 35(1):11~18
14 S.Amari. The Present and Future of Neural Network. Kyoritsu Press. 1990, inJapanese
15 P. Moerland. A Comparison of Mixture Models for Density Estimation Artifical Neural Networks. Conference Publication. 1999,(470):7~10
16 F.Takeda, Sigeru Omatu. High Speed Paper Currency Recognition by Neural Networks. IEEE Trans Neural Network. 1995, l6(1)
17 Frosini, M. Gori, P. Priami. A Neural Network-Based Model for Paper Currency Recognition and Verification. IEEE Transactions on Neural Networks. 1996, 7(6):1482~1490
18李继飞.神经网络模式识别技术在自动售货机中的应用研究. 湖南大学硕士学位论文. 2001
19 F. Takeda, S. Omatu. High speed paper currency recognition by neural networks. IEEE Trans on Neural Network. 1995, 6 (1):73~77
20 F. Takeda, S. Omatu, T. Inoue, and S. Onami. A Structure Reduction of Neural Network with Random Masks and Bill Money Recognition. Proceedings of the 2nd International Conference on Fuzzy Logic and Neural Networks. 1999, 12(5):
21 H.D. Cheng, Y.Sun. A hierarchical approach to color image segmentation using homogeneity. IEEE Trans on Image Processing. 2000, 9(12):2071~2082
22 TMS320C6713B FLOATING-PINT DIGITAL SIGNAL PROCESSOR
23 TMS320C6000 ON CHIP SUPPORT LIBRARY API USER’S GUIDE
24 TMS320C6X Peripheral Support Library Programmer’s Guide
25 TMS320C6000 McBSP: UART
26 TMS320C6000 External Memory Interface Reference Guide
27 T. Michael, U.E. Martin. Digital Processor for Wavelet-based Image Compression. Proceedings of SPIE - The International Society for Optical Engineering. 2000, (3965):340~347
28徐婉莹,刘建军,黄新生.基于CPLD和DSP的高速图像采集技术研究. 电子工程师. 2004, 30(06):48~50
29 Y.L. Li, G.Q. Ni, Chen Xiao-Mei. Control for Image Fusion at Real Time Implemented with Programmable Devices. Opto-Electronic Engineering. 2003, 30(4):11~14
30 Complete 14-bit CCD/CIS Signal Processor
31杨洁,冯力刚,蒋加伏. 基于小波包和支持向量机的人脸识别. 2004, 21(9): 131~133
32 Cygnal Integrated Products,INC.潘琢金,孙德龙,夏秀峰译. C8051F单片机应用解析.北京航空航天大学出版社. 2002,10
33崔明星.纸币识别算法的研究.哈尔滨工业大学硕士学位论文. 2002
34刘家锋, 刘松波, 唐降龙.一种实时纸币识别方法的研究. 计算机研究与发展. 2003,40(7): 1057~1061
35 M. Moerland. A Comparision of Mixture Models for Density Estimation Artificial Neural Networks. Conference Publication IEE. 1999,(470):7-10
36 U.V. Chaudhari, J. Navratil. Multigrained Modeling with Pattern Specific Maximum Likelihood Transformations for Text-Independent Speaker Recognition IEEE Transactions on Speech and Audio Processing. 2003, 11(1):61~69
37 Bhattacharyya, T. Srikanthan, P. Krishnamurthy. Ideal GMM Parameters & Posteriorlog Likelihood for Speaker Verification. IEEE 2001, 3(5):208~213
38 Hedelin , J. Skoglund. Vector Quantization Based on Gaussian Mixture Models IEEE Trans. Speech and Audio Processing. 2000, 8(4):385~401
39 A. Dempster, N. Laird, D. Rubin. Maximum Likelihood from Incomplete Data via the EM Algorithm, J. Roy. Stat. Soc. 1977, 39(3):1~38
40刘关立. 任意文本的说话人识别系统研究. 大连理工大学硕士论文 2000.3
41 Dempster, A.P.Laird, N.M. and Rubin. Maximum-Lilelihood from Incomplete Data Via The EM Algorithm Journal of Royal Statistics cs Society Series. 1977, 99(1):1~38
42 J.A.Bilmes. A Gentle Tutorial of the EM Algorithm and its Application to Parameter Estimation for Gaussian Mixed and Hidden Markov Models U.C.Berkely, TR-97-021, 1998
43 Fumiaki Takeda, Sigeru Omatu. High speed paper currency recognition by neural networks. IEEE Trans on Neural Network. 1995, 6(1):73~77
44 A Frosini, M Gori, P Priami. A neural network based model for paper currency recognition and verification. IEEE Trans on Neural Networks. 1996, 7(6): 1482~1490
45 J.F. Liu, S.B. Liu, and X.L. Tang. An Algorithm of Real Time Paper Currency Recognition. Journal of computer research and development. 2003, 140(17):1057～1061
46 A. Latif-Amet, A. Ertu’zu’n, A. Erc?il. An efficient method for texture defectdetection: sub-band domain co-occurrence matrices. Image and Vision Computing . 2000, (18):543~553
47 H. Onishi, S. Tatsumi. A Pattern Defeat Inspection Method by Parallel Grayscale Image Comparison without Precise Image Alignment. IEEE 28th Annual Conference of the Industrial Electronics Society. 2002, 3(5-8): 2208~2213
48 L.G. Brown. A survey of image registration techniques. ACM Computing Surveys. 1992, 24(4):325~376
49 B. Zitová, J. Flusser. Image registration methods: a survey. Image and Vision Computing. 2003, 21(11):977~1000