臺灣博碩士論文加值系統

自動車牌識別在自動化和智能交通系統管理上扮演非常重要的作用。目前的車牌識別系統只能識別單目標，無法進行多目標識別。在實際應用面，相機拍攝的圖像可能具有多個目標，單目標識別會放棄很多信息降低效率。為了改善這種車牌識別系統，本論文提出一種新的模板匹配方法以及改進一些影像處理技術供車牌識別，然後整合起來既有建立一個高精度多個車牌識別系統。
本論文所提出之高精度多個車牌識別系統分為預先處理，車牌區域分離，字元分割，以及字元識別等四部分。首先，預處理部分使用中值濾波器，邊緣檢測，數學形態學和連通域分析法找出可能的車牌區域。然後，在可能區域內我們使用基於字元特徵垂直投影法來確認車牌區域。在字元分割部分，本文採用之方法是以投影法為基礎，最後，字元識別是採用本文所提出的新模板匹配方法。
實驗結果顯示，我們所提出的系統可以在一張有複雜的背景影像中有效地定位以及識別多個車牌。整體實驗結果顯示，所提出的系統對於單目標辨識準確率達91.5％，以及對於多目標的辨識準確率為90.9％。

Automatic license plate recognition plays a very important role in the management of automation and intelligent transportation systems. However, the license plate recognition systems at the present day can only recognize a single target. In practice applications, an image captured from the camera may contain multiple targets. Many pieces of information will be left behind if the single-target license plate recognition system is applied to such a picture. To reform the license plate recognition system, a new template matching method is proposed and several image processing schemes for character recognition are modified. As a consequence of integrating the proposed and reforming schemes, a multiple-target license plate recognition system of high precision is developed and established.
The multiple-target license plate recognition system is comprised of preprocessing, plate-region detection, character segmentation, and character recognition. First in the preprocessing part, the median filter is used. The edge detection, mathematical morphology, and connected component labeling methods are used in sequence to find out possible regions of license plates. Then the vertical projection method based on characters’ features is applied in the ranges of possible regions for ensure the regions of license plates. The projection method is used for character segmentation then, and the proposed template matching scheme is used for character recognition
Experimental results show that the proposed template matching method is robust in extracting and recognizing multiple license plates from the image with a complex background. Our experiments also show that the proposed license plate recognition system has 91.5% of the accurate rate for a single target recognition and 90.9% of the accurate rate for multi-target recognition.

Abstract in Chinese..........................................................................................I
Abstract in English..........................................................................................II
致謝..................................................................................................................IV
Contents............................................................................................................V
List of Figures ……..........................................................................................VIII
List of Tables…….............................................................................................XI
Chapter 1 Introduction .................................................................................... 1
Chapter 2 Literature Review and Related Technology……………..........…3
2.1 Literature Review …………….............................................................. 3
2.1.1 License Plate Localization…………………………………..…...3
2.1.2 Character Segmentation…………………………………….……4
2.1.3 Character Recognition……………………………………...…....5
2.2 Vehicle License Specifications in Taiwan..............................................6
2.3 Related Technology................................................................................7
2.3.1 Median Filter……………………………………………………..7
2.3.2 Edge Detection..............................................................................8
2.3.3 Binarization...................................................................................9
2.3.4 Morphological Processing............................................................12
2.3.5 Connectivity…………………….................................................13
2.3.6 Normalization..............................................................................16
2.3.7 Projection Method………………………………………….…..17
Chapter 3 Framework of License Plate Recognition System………….…..19
3.1 The Pre-processing Procedure..............................................................20
3.1.1 Color To Gray Images................................................................ 21
3.1.2 Median Filter………………………………………………...….21
3.1.3 Sobel Edge Detection……………………………………….…..23
3.1.4 Image Binarization………………………………………….…..24
3.1.5 Morphological Processing……………………………….….….25
3.2 Detection Procedure of License Plate…………………………………28
3.3 Character Segmentation Procedure…………………………………...36
3.3.1 Binarization Invers……………………………………..….……38
3.3.2 Projection Method……………………………………………...39
3.3.3 Connected Component…………………………………..……..41
3.3.4 Angle Calibration……………………………………………….42
3.3.5 Character Segmentation…………………………………………44
3.4 Character Recognition Procedure..........................................................45
3.4.1 Character Normalization.............................................................46
3.4.2 Character Database.....................................................................46
3.4.3 Character Recognition.................................................................48
Chapter 4 Experiment Results and Discussion............................................50
4.1 Experiments..........................................................................................50
4.2 Results..................................................................................................51
4.2.1 License Plate Detection…………………………………….……52
(1) Single Target..........................................................................52
(2) Multi-Target...........................................................................53
4.2.2 Character Segmentation……………………………….…………53
(1) Single Target...........................................................................53
(2) Multi-Target............................................................................53
4.2.3 Character Recognition…………………………………………...54
(1) Single Target...........................................................................54
(2) Multi-Target............................................................................54
4.3 Causes of Incorrect Detection for License Plates………………………54
4.4 Discussion...............................................................................................56
Chapter 5 Conclusions……………………………………………..……..……57
References...........................................................................................................59

[1] JieHan Yu, Jianqi An and Gang Xiong, “A license plate recognition system based on machine vision,” IEEE International Conference. , pp. 259-263, July 2013
[2] Qing Tian, Teng Guo, Long Zhang and Shuai Qiao,“Research of automatic parking system based on license plate recognition,” IEEE International Conference. , pp.363-366, Nov. 2013
[3] A.B.Mello Carlos and C.Costa Diogo,“A complete system for vehicle license plate recognition,” Signals and Image Processing, 2009. 16th International Conference, no. 2, pp. 1-4, June 2009
[4] Megalingam, P. Krishna, P.Somarajan, V.A Pillai and R.U Hakkim, “Extraction of license plate region in automatic license plate recognition,” Mechanical and Electrical Technology (ICMET), 2010 2nd International Conference, no. 1, pp. 496-501, Sept. 2010
[5] Lu Liu, Hongjiang Yu, Kehe Cai and Jia Wang, “License plate recognition using topology structure features,” Control and Industrial Engineering (CCIE), 2011 IEEE 2nd International Conference, vol. 2, pp. 251-254, Aug. 2011
[6] Haiyan Zhao, Chuyi Song, Haili Zhao and Shizheng Zhang, “License plate recognition system based on morphology and LS-SVM,” 2008 International Conference, no. 1, pp. 826-829, Aug. 2008
[7] Bo Li, Zhiyuan Zeng, Jianzhong Zhou and Huali Dong, “An algorithm for license plate recognition using radial basis function neural network,” Computer Science and Computational Technology 2008 International Symposium, vol. 1, no. 4, pp. 569-572, Dec. 2008
[8] C. Wu, L. C. On, C. H. Weng, T. S. Kuan, and K. Ng, “A macao license plate recognition system,” in Proc. IEEE 4th Int. Conf. , Guangzhou, China, vol. 7, no. 2, pp. 4506–4510, Aug. 2005
[9] K. Deb, M.K. Hossen, M.I. Khan, and M.R. Alam “Bangladeshi vehicle license plate detection method based on HSI color model and geometrical properties,” 2012 7th International Forum, pp. 1-5, 2013
[10] L. Dlagnekovin, “License plate detection using adaboost,” 2004 Comput. Sci. Eng. Dept., Univ. California San Diego, Mar. 2004
[11] N.Vishwanath, S.Somasundaram, M.R.Rupesh Ravi, and N.Krishnan Nallaperumal, “Connected component analysis for Indian license plate infra-red and color image character segmentation,” 2012 IEEE International Conference on, pp. 1-4, Dec. 2012
[12] Xin Fan, Guoliang Fan and Dequn Liang, “Joint segmentation and recognition of license plate characters,” Image Processing, 2007 IEEE International Conference, vol. 4, no. 2, pp. 353-356, Oct. 2007
[13] Xin Fan and Guoliang Fan, “Graphical models for joint segmentation and recognition of license plate characters,” IEEE Signal Processing Letters, vol. 16, no. 1, January 2009
[14] R. Huang, H. Tawfik, and A.K. Nagar, “License plate character recognition based on support vector machines with clonal selection and fish swarm algorithms,” Computer Modelling and Simulation, 2009.IEEE 11th International Conference on. , pp. 101-106, March 2009
[15] K. Omar, S. Sahran, and M. Khalid, “License plate recognition based on support vector machine,” Electrical Engineering and Informatics, ICEEI 2009 International Conference on, vol. 1, pp. 78-82, Aug. 2009
[16] Wei Zhu, Mei Xie, and Jian-feng Xie, “A decision tree algorithm for license plate recognition based on bagging,” Wavelet Active Media Technology and Information Processing (ICWAMTIP), 2012 International Conference on, vol. , no. , pp. 136-139, Dec. 2012
[17] E. Anagnostopoulos Christos Nikolaos, and E. Anagnostopoulos Ioannis, “A license plate recognition algorithm for intelligent tranportaion system applications,” IEEE Transactions on Intelligent Transportation Systems, vol. 7, no. 3, September 2006
[18] Yanping Hu, Hongpeng He, and Decheng Ma, “Research on recognition of license plate in relative moved environment,” Journal of Dalian University of Technology, 44(3), pp. 412-415, 2004
[19] Shyang-Lih Chang, Li-Shien Chen, Yun-Chung Chung, and Sei-Wan Chen, “Automatic license plate recognition,” IEEE Transactions on Intelligent Transportation Systems. , vol. 5, no. 1, March 2004
[20] Fu-guo Dong, “A novel fast algorithm for image median filtering based on improved sign test method,” Logistics Systems and Intelligent Management, 2010 International Conference on, vol. 2, no. 1, pp. 1240-1242, Jan. 2010
[21] Nobuyuki Otsu, “A threshold selection method from gray-level histograms”, IEEE Transactions on Systems, man, and cybernetics, vol. 9, no. 1, 1979
[22] Xiaohui Mao and Liying Zheng, “Locating micrometer in a digital image based on projection algorithm in space domain,” Internet Computing for Science and Engineering (ICICSE), 2010 Fifth International Conference on, pp. 30-33, Nov. 2010