臺灣博碩士論文加值系統

中文摘要
雖然車牌辨識系統已有多年的研究，在車牌定位處理的過程中仍會佔用許多時間，因此發展一個快速的車牌定位系統是非常重要。差分法是最簡單的邊緣偵測方法，運算過程少，可是效果卻不理想，因為會有其他影像邊緣會影響車牌定位。在本論文中，將差分法做了一些改善，使車牌位置在影像中可以更容易被定位。本研究主要包括車牌定位、二值化、車牌字元的分割及字元的辨識。在車牌定位的部份，使用水平移位差分法加強影像車牌文字邊緣之效果，使我們可以用簡單投影法在影像中找出車牌的位置。使用垂直軸投影法先找出車牌的高，再利用水平軸投影法找出車牌的寬度。水平方向的車牌文字中有空白的部份，這些空白的部份會將車牌不當的分割，我們可以利用平滑化的方式將分割點去除找出正確的車牌位置。找出的車牌影像利用二值化的方法將字元及背景分離，為了可以在陰影中將車牌影像及文字分離，我們使用動態二值化的方法，它可以隨著不同的環境改變二值化的值，使車牌字元可以清楚的顯示。最後，利用部份辨識法辨識所分割之字元影像，找出影像中的車牌號碼。
本實驗中使用數位相機到室內及室外停車場拍攝180張車牌影像，機車部份為80張，汽車部份為100張，其中車牌的定位率可達98.8%、辨識率可達92.7%，每張影像的平均辨識時間只需要0.586秒，大大的改善了程式執行的時間。

ABSTRACT
Differencial method is the simplest edge detection method. The alogrithm process is fewer but has a bad result because there are some other image edge would affect the license plate location. In this thesis, we make some improvement in the differential method to make the plate position easier to be located in the image. The recognition system includes the license plate location, binarization, the segmentation of the license plate words and the word recognition. In the license plate location, we use horizontal shifty differential method to reinforce the effect of the license plate word edge, and then we can use the simple projection to find the position of the license plate. We use the vertical axel projection to find out the height of the license plate and use the horizontal axel projection to find out the width of the license plate. There are some blank parts in the horizontal license plate words. These blanks would segment the license plate incorrectly. We could use the modification to get rid of the division points and find out the correct license plate position. We use the binarization to segment the word and the background of the license plate image. In order to segment the license image and the word in the shadow, we use the dynamic binarization which could change its value according to the different environments and make the license plate words show clearly. We then use the vertical and horizontal projection to segment the license plate images. At last, we use the partial recognition method to recognize the segment word images and then we can find out the license plate number in the image.
There are 80 motercycles images and 100 automobiles images as the experimental samples. With the situation, it is found that the rate of the license location could achieve 98%, and the rate of the recognition could achieve 92.7%. The average recognition time of each image only needs 0.586 second and it improves the time of the program execution a lot.

TABLE OF CONTENTS
ENGLISH ABSTRACT i
CHINESE ABSTRACT ii
ACKNOWLEDGEMENTS iii
TABLE OF CONTENTS iv
LIST OF FIGURES vi
LIST OF TABLES x
CHAPTER
I INTRODUCTION 1
1.1 Motivation 1
1.2 Objectives 3
1.3 Thesis Organization 3
II BACKGROUND 5
2.1 Related Researches 5
2.2 System Organizations 9
2.2.1 Hardware 9
2.2.2 Software 9
2.2.2.1 License Palte Locating 10
2.2.2.2 Binarization and Segmentation 12
2.2.2.3 Word Recognition 13
III LICENSE PLATE RECOGNITION SYSTEM 15
3.1 Normalization 15
3.2 Shifty Differential Method 15
3.3 Binarization 34
3.4 Word Segmentation 46
3.5 Word Recognition 48
IV EXPERIMENTAL RESULTS 54
4.1 Locating License Plate 54
4.2 Binarization and Word Segmentation 59
4.3 Word Recognition 62
4.4 Results 63
V CONCLUSIONS AND FUTURE WORKS 64
5.1 Conclusions 64
5.2 Future Works 66
REFERENCES 68

REFERENCES
[1] Y. L. Chang, An Automatic and Real-Time Recognition for Vehicle License Plate System by Systemic Approch, Master Thesis, Department of Computer Science and Information Engineering, Chung Hua University, 2002.
[2] J. Y. Liao, A System for the Automatic and Real-Time Recognition of Vehicle License Plate, Master Thesis, Department of Computer Science and Information Engineering, Chung Hua University, 2000.
[3] Z. Y. Li, The Study of Car License Recognition System, Master Thesis, Department of Information Science and Management with MBA, 2003.
[4] H. T. Lue, Recognition System of License Plate Using Multi-Experts, Master Thesis, Graduate Institude of Computer Science and Information Engineering, National Central University, 2002.
[5] S. Y. Lai, The Design and Application of Template-Base License Plate Recognition system, Master Thesis, Graduate Institude of Conputer Science Engineering, Tatung University, 2002.
[6] J. L. Chen, Automatic License Number Recognition for Car Image Taken for Parking Violation, Master Thesis, Institude of Computer and Information Science, National Chiao Tung University, 2001.
[7] C. W. Lee, Vehicle Plate Recognition System, Master Thesis, Department of Electrical Engineering, National Taiwan University of Science and Technology, 2000.
[8] J. B. Fang, A Study of Car-Plate Number Recognition System, Master Thesis, Department of Engineering Science, National Cheng Kung University, 1999.
[9] Y. C. Wei, Motion Based Multiple Vehicle License Plate Recognition, Master Thesis, Graduate Institude of Information Management, Yuan Ze University, 2000.
[10] S. P. Lin, The Extraction of Vihicle Plate Characters, Master Thesis, Department of Electrical and Control Engineering, National Chiao Tung University, 1998.
[11] J. H. Juang, The Recognition of Vehicle Plate Characters Using WMMS Method, Master Thesis, Department of Electrical and Control Engineering, National Chiao Tung University, 1999.
[12] J. J. Tash, Character Partition Normalization and Recognition Resarch of License Plate Recognition System, Master Thesis, Department of Electrical and Control Engineering, National Chiao Tung University, 1999.
[13] H. H. Lin, C. Y. Chen and J. H. Chuang, “Recognition of Printed Digits of Low Resolution,” Pattern Recognition and Image Analysis, vol.10, no. 2, pp. 265-272, Dec. 2000.
[14] P. Comelli, P. Ferragina, M. N. Granieri and F. Stabile, “Optical Recognition of Motor Vehicle License Plates,” IEEE Trans. on Vehicular Technology, vol.44, no. 4, Nov. 1995.
[15] T. Naito, T. Tsukada, K. Yamada, K. Kozuka and S. Yamamoto, “Robust License-Plate Recognition Method for Passing Vehicles under Outside Environment,” IEEE Trans. on Vehicular Technology, vol. 49, no. 6, Nov. 2000.
[16] N. Otsu, “A Threshold Selection Method from Gray-Level Histograms,” IEEE Trans. on System, Man, and Cybernetics, vol. SMC-9, pp. 62-66, no. 1, Jan. 1979.
[17] X.F. Hermida, F.M.Rodriguez, J.L.F Lijo, F. P. Sande, and M. P. Iglesias, “A System for the Automatic and Real Time Recognition of V. L. P.’s (Vehicle License Plate),” Lecture Note in Computer Science, Vol. 1311, pp. 552-558, Jun. 1997.
[18] R. S. Duh, Edge Detection and its Performance Evaluation, Master Thesis, Institute of Computer and Decision Sciences, National Tsing Hua University, 1986.
[19] Y. T. Zhuang, Text Detection in Color Iimages and Compound Document Compression, Master Thesis, The Graduate Institute of Communacation Engineering, National Taiwan University, 2003.
[20] L. Jianzhuang, L. Wenqing, and T. Yupeng, “Automatic Threshoding of Gray-level Pictures Using Two-Dimension Otsu Method,” International Conf. on Circuits and Systems, vol. 1, pp. 325-327, June 1991.
[21] G. Ramponi, “Edge Extraction by a Class of Second-order Nonlinear Filters,” Electron. Lett., Vol. 22, no. 9, Apr. 1986.
[22] A. Huertas and G. Medioni, “Detection of Intensity Changes with Subpixel Accuracy Using Laplacian Gaussian Masks,” IEEE Trans. Pattern Anal. Mach. Intelligence, vol. PAMI-8, no. 5, pp. 651-664, 1986.
[23] R. Hummel, “Reconstructions form Zero Crossing in Scale Space,” IEEE Trans. Pattern Anal. Mach. Intelligence, pp. 2111-2130, 1989.
[24] J. H. Liu, “A Research on Designing an Adaptive Image Edge Detection System with Statistical Method,” Master Thesis, Department and Graduate Institude of Automatic Control Engineering, Feng Chia University, 1989.
[25] Larry S. Davis, “A Survey of Edge Detection Techniques,” Computer Graghic and Image Processing, vol. 4, pp. 248-270, 1975.