臺灣博碩士論文加值系統

近年來車輛俱以日增，為了有效控管車輛，我們研發一套車牌辨識系統可應用於停車管控、贓車追緝與車輛檢驗等範疇。本篇論文提出一套低運算量、高辨識率的車牌辨識流程，並實現了在ARM平台上車牌辨識系統的開發與規劃，以節省成本，同時達到即時偵測的效能。
車牌辨識系統主要有三個階段：1. 車牌偵測、2. 字元分割、3. 字元辨識，我們提出一個以小波為基礎的車牌偵測，利用紋理特徵、投影量、群組化的概念將車牌擷取出來；經過區域二值化後，再使用連通物件連接法將字元作分割，取得個別字元影像，先做正規化處理運算，再利用樣板比對法將字元辨識出來。
利用上述方法，我們可以在ARM平台上偵測到車牌並成功切割出字元，在不設限外在環境和影像中車牌位置的情形下，即使原始影像有些許傾斜或有不同強度光線照射，經實驗證實，我們可以得到約95.45%的成功偵測率。如果在既定條件下，距離車體1~3公尺、俯角25~35度、車牌傾斜不超過10度、車牌無明顯髒汙毀損，我們可以百分之百地成功擷取出車牌而字元切割成功率也達到97.73%；單一字元辨識率則有94.05%，若在字元乾淨且車牌無變形的情況下，我們亦可百分之百的辨識出個別字元。

A license plate recognition system is composed of three processing modules: license plate detection, character segmentation, and character recognition. This paper presents a method of vehicle license plate image detection based on wavelet transform, character segmentation using connected components labeling, and recognition based on template matching.
The proposed system consists of four main stages. First, we convert images to 8-bit grayscale and adjust the images to proper size. And then, a wavelet transform based method is used to get important contrast features as guides to search for desired license plates. Second, finding a reference line in LH sub-image plays an important role to locate the desired license plate region. According to the reference line we can decrease the searching region of license plate and speed up the execution time. Third, we binarize our image by using Otsu algorithm. After that, we use 4-connected components labeling algorithm to segment characters of the license plates. Forth, we normalize the size of the segmented characters and recognize characters using template matching method.
Since the feature extracted is robust to complex backgrounds. The major advantage of the proposed method is that it works well in extracting differently illuminated and oriented license plates.
We experiment with about one hundred vehicle images obtained from basement parking area. The average accuracy of detection is 95.45% and the average accuracy of segmentation is 96.59%. When we take the pictures with angle of depression 25~35 degrees and 1~3 meter away from the vehicle without any shelters or dust on the license plates, we will receive 100% accuracy of the detection and segmentation result. However, the license plate can not slope over 10 degrees. The average recognition result is 94.05%.

論文摘要 i
ABSTRACT ii
誌謝 iii
目錄 iv
圖索引 vi
表索引 viii
第一章 緒論 - 1 -
1.1 前言 - 1 -
1.2 研究動機 - 2 -
1.3 研究背景 - 3 -
1.4 論文架構 - 5 -
第二章 相關研究 - 6 -
2.1 常見車牌偵測方法 - 6 -
2.1.1 紋理特徵 - 6 -
2.1.2 色彩訊號 - 7 -
2.2 二值化 (Binarization) - 7 -
2.2.1 固定門檻值 - 8 -
2.2.2 平均像素門檻值 - 9 -
2.2.3 統計式門檻值 - 10 -
2.2.4 Entropy門檻值 - 12 -
2.3 常見的字元分割方法 - 13 -
2.3.1 投影量分割法 (Projection Histogram) - 14 -
2.3.2 外圍輪廓值分割法 - 14 -
2.3.3 連通物件標示法 (Connect Component Labeling) - 15 -
2.4 本系統架構 - 16 -
第三章 車牌辨識系統 - 18 -
3.1 影像前置處理 - 18 -
3.2 離散小波轉換 (Discrete Wavelet Transform) - 20 -
3.2.1 基本原理 - 20 -
3.2.2 二維離散小波轉換 - 21 -
3.2.3 Haar轉換 - 25 -
3.2.4 提升式結構 (Lifting Scheme) 小波轉換 - 29 -
3.3 車牌擷取 - 33 -
3.3.1 影像參考線 - 35 -
3.3.2 車牌偵測找尋 - 36 -
3.4 字元分割 - 39 -
3.4.1 區域二值化 - 41 -
3.4.2 區塊化 - 41 -
3.4.3 字元區塊調整 - 45 -
3.5 字元辨識 - 47 -
3.5.1 字元正規化（Character Normalization） - 49 -
3.5.2 樣板比對法 - 51 -
第四章 系統測試與結果 - 53 -
4.1.1 實作資料 - 55 -
4.1.2 實作結果 - 56 -
第五章 結論與未來展望 - 68 -
5.1 結論 - 68 -
5.2 未來展望 - 68 -
參考文獻 - 69 -

參考文獻
[1]楊汝浩，汽車牌照識別，國立交通大學控制工程研究所，民國75年。
[2]S.-Z. Wang and H.-J. Lee, “Detection and recognition of license plate characters with different appearances,” Intelligent Transportation Systems, 2003. Proceedings. 2003 IEEE, vol. 2, 2003, pp. 979-984.
[3]S. Kim, D. Kim, Y. Ryu, and G. Kim, “A robust license-plate extraction method under complex image conditions,” Pattern Recognition, 2002. Proceedings. 16th International Conference on, vol. 3, 2002, pp. 216-219.
[4]H. Bai, J. Zhu, and C. Liu, “A fast license plate extraction method on complex background,” Intelligent Transportation Systems, 2003. Proceedings. 2003 IEEE, vol. 2, 2003, pp. 985-987.
[5]H. Bai and C. Liu, “A hybrid license plate extraction method based on edge statistics and morphology,” Pattern Recognition, 2004. ICPR 2004. Proceedings of the 17th International Conference on, vol. 2, 2004, pp. 831-834.
[6]V. Kamat and S. Ganesan, “An efficient implementation of the Hough transform for detecting vehicle license plates using DSP’S,” Real-Time Technology and Applications Symposium, 1995. Proceedings, 1995, pp. 58-59.
[7]S. Gendy, C. L. Smith, and S. Lachowicz, “Automatic car registration plate recognition using fast Hough transform,” Security Technology, 1997. Proceedings. The Institute of Electrical and Electronics Engineers 31st Annual 1997 International Carnahan Conference on, 1997, pp. 209-218.
[8]Jianlong Zhu, Yannan Zhao, “Vehicle license image segmentation using wavelet transform,” Intelligent Multimedia, Video and Speech Processing, 2001. Proceedings of 2001 International Symposium on, 2001, pp. 267–270.
[9]Ching-Tang Hsieh, Yu-Shan Juan, Kuo-Ming Hung, “Multiple license plate detection for complex background,” Advanced Information Networking and Applications, 2005. AINA 2005. 19th International Conference on, vol. 2, 2005, pp. 389-392.
[10]W. Wei, M. Wang, and Z. Huang, “An automatic method of location for number-plate using color features,” Image Processing, 2001. Proceedings. 2001 International Conference on, vol. 1, 2001, pp. 782-785.
[11]S. H. Park, K. I. Kim, K. Jung, and H. J. Kim, “Locating car license plates using neural networks,” Electronics Letters, vol. 35, 1999, pp. 1475-1477.
[12]S. K. Kim, D. W. Kim, and H. J. Kim, “A recognition of vehicle license plate using a genetic algorithm based segmentation,” Image Processing, 1996. Proceedings, International Conference on, vol. 1, 1996, pp. 661-664.
[13]Shyang-Lih Chang, Li-Shien Chen, Yun-Chung Chung, and Sei-Wan Chen, “Automatic license plate recognition,” Intelligent Transportation Systems, IEEE Transactions on, vol. 5, Issue 1, 2004, pp. 42-53.
[14]林泰良，智慧型車牌定位與字串分割，國立台灣大學電機工程所，2000。
[15]N. Otsu, “A Threshold Selection Method from Gray-Level Historgams,” IEEE Transactions on System, Man, and Cybernetics, vol. SMC-9, 1979, pp. 62-66.
[16]D.Y. Xue, V. Gupta, “An improved threshold selection method for image segmentation”, Electrical and Computer Engineering, 1993, pp 531-534.
[17]J. N. Kapur, P. K. Sahoo, and A. K. C. Wong, “A new method for gray-level picture thresholding using the entropy of the histogram,” Computer Vision, Graphics, and Image Processing, 1985, pp. 273-285.
[18]呂炎州，不需字元切割的車牌辨識法，靜宜大學資訊管理學系研究所，2003。
[19]Y. Lu, “On the Segmentation of Touching Characters,” Document Analysis and Recognition, 1993, Proceedings of the Second International Conference on, 1993, pp. 440-443.
[20]官宗，利用數位訊號處理器實現車牌字元辨識系統，國立台灣大學電機工程學研究所，民國90年。
[21]Rafael C. Gonzalez and Richard E. Woods, Digital Image Processing, 2nd Edition, Prentice-Hall, New Jersey, 2002
[22]吳忠政，二維三階上提式離散小波轉換架構，南台科技大學碩士論文，2002
[23]Sweldens, “The Lifting Scheme: A Custom Design and Contruction of Biorthogonal Wavelets”, Appiled and Computational Harmonic Analysis (ACHA), pp.186-200, 1996
[24]戴顯權、陳政一，JPEG2000，紳藍出版社，2002
[25]Ching-Tang Hsieh, Yu-Shan Juan, Kuo-Ming Hung, “Multiple license plate detection for complex background,” Advanced Information Networking and Applications, 2005. AINA 2005. 19th International Conference on, vol. 2, 2005, pp. 389-392.
[26]林陽彬、陳同孝、王清德，應用多頻道韋納過濾技術辨識模糊車牌字元影像，國立台中技術學院，2004
[27]P. Comelli, P. Ferragina, M. N. Granieri, and F. Stabile, “Optical recognition of Motor Vehicle License Plates” IEEE Trans. on Vehicular Tech, vol. 44, no.4, November 1995.