臺灣博碩士論文加值系統

現今許多車牌辨識方面的研究都是使用固定式攝影機與個人電腦來達成，然而這種系統大多會設下許多限制條件，如:固定背景、特定拍攝角度和畫面中車輛數，因此在我們研究裡會盡可能的減少辨識環境的限制。
所以本論文將探討一個嵌入式系統，畫面中能同時存在多輛汽機車，且容許較大的拍攝角度，並使用較低運算複雜度的演算法。此系統共包含四個區塊：車牌定位、歪斜車牌矯正、字元切割、字元辨識，一開始我們先利用車牌對比明顯的特徵來找出畫面中可能的車牌，之後再利用影像幾何轉換來校正車牌，接著使用投影法切出車牌上的字元，以較具有適應性的樣板比對來完成車牌的辨識。
最後將我們提出之演算法實現於SIMIS BA-8嵌入式平台上，此系統的優點是有多車牌辨識能力、方便攜帶且擁有大角度拍攝辨識的能力。實驗結果顯示此架構有相當靈活性且能夠達到每秒八張以上的即時辨識速度，並在大部份的環境中都有不錯之辨識率。

Nowadays, most researches on the license plate recognition system are implemented by computers with fixed cameras. However, most of them operate under restricted conditions, such as static background, fixed shooting angle, and limited number of cars. In this research, as few constraints as possible on the working environment are considered.
Our study focuses on the development of a real-time multiple vehicles license plate recognition system on embedded devices with algorithms in low computational complexity and camera shooting angle in large range. It consists of four processing modules: plate localization, skew correction, characters segmentation, and character recognition. By the characteristics of high-contrast of plate image, the license plates are first localized. After detecting the skew lines, the skew plate images are corrected by geometric transformation. Then, applying projection method, the characters on each plate are segmented. Finally, a recognition algorithm by an adaptive template matching is applied to identify the characters on the license plate.
The proposed algorithms are implemented on SIMIS BA-8 board. The major advantages of our system are that it accomplishes a multiple license plate detection and recognition system and it is a portable real-time device with large range in shooting angle.
Our algorithm was tested with video streams in different shooting angles, distances, complex background, and variant illumination. The experimental results show that our proposed architecture is a flexible system. It can operate in real-time at a frame rate of 8 fps with good recognition rate in most circumstances.

摘要 I
Abstract II
誌謝 III
目錄 IV
圖目錄 VI
表目錄 VIII
第一章 緒論 1
1.1 研究動機與背景 1
1.2 文獻回顧 2
1.2.1 車牌定位文獻探討 2
1.2.2 車牌校正文獻探討 3
1.2.3 字元切割文獻探討 3
1.2.4 字元辨識文獻探討 4
1.3 論文目標 4
1.4 論文組織 5
第二章 系統流程與實驗環境 6
2.1 系統流程 6
2.2 實驗環境與驗證平台 7
2.2.1 實驗環境 7
2.2.2 驗證平台 8
第三章 車牌辨識系統 11
3.1 車牌定位 11
3.1.1 邊緣偵測 12
3.1.2 水平連線 16
3.1.3 侵蝕膨脹 16
3.1.4 物件連通標記法 19
3.1.5 計算特徵值 23
3.1.6 分析車牌顏色 25
3.2 車牌校正 26
3.2.1 最小方差法 27
3.2.2 邊框提取 29
3.2.3 霍夫轉換法測線 30
3.2.4 影像幾何轉換 33
3.3 字元切割 37
3.3.1 平均二值化 38
3.3.2 去除非車牌區域 39
3.3.3 面積法二值化 39
3.3.4 影像投影法字元切割 41
3.3.5 物件連通標記法字元切割 44
3.4 字元辨識 46
3.4.1 Necessity-sufficiency matching 47
3.4.2 相似字元判斷 50
第四章 實驗結果與效能分析 52
4.1 多車牌不同距離實驗結果 53
4.2 多車牌不同角度實驗結果 55
4.3 低光源實驗結果 57
4.4 模擬實際使用狀況實驗結果 59
4.5 各部份成功率比較 61
4.6 系統在嵌入式平台執行時間分析 61
第五章 結論與未來研究方向 63
5.1 結論 63
5.2 未來研究方向 63
參考文獻 64
作者簡介 66

[1]J. M. Guo and Y. F. Liu, “License plate localization and character segmentation with feedback self-learning and hybrid binarization techniques,” IEEE Transactions on Vehicular Technology, vol. 57, no. 3, pp. 1417-1424, May. 2008.

[2]J. Q. Song, Z. Li, M. R. Lyu, and S. J. Cai, “Recognition of merged characters based on forepart prediction, necessity-sufficiency matching, and character-adaptive masking,” IEEE Transactions on Systems, Man, and Cybernetics, Part B-Cybernetics, vol. 35, no. 1, pp. 2-11, Feb. 2005.

[3]T. D. Duan, T. L. H. Du, T. V. Phuoc, and N. V. Hoang, “Building an automatic vehicle license-plate recognition system,” Proceedings of the International Conference in Computer Science RIVF'05, pp. 59-63, Feb. 2005.

[4]M. Yu and Y. D. Kim, “An approach to Korean license plate recognition based on vertical edge matching,” Proceedings of the IEEE International Conference on Systems, Man, and Cybernetics, vol. 4, pp. 2975-2980, 2000.

[5]M. Sarfraz, M. J. Ahmed, and S. A. Ghazi, “Saudi Arabian license plate recognition system,” Proceedings of the International Conference onGeometric Modeling and Graphics, pp. 36-41, Jul. 2003.

[6]C. Wu, L. C. On, C. H. Weng, T. S. Kuan, and K. Ng, “A Macao license plate recognition system,” Proceedings of the Fourth International Conference on Machine Learning and Cybernetics, pp. 4506-4510, Aug. 2005.

[7]S. L. Chang, L. S. Chen, Y. C. Chung, and S. W. Chen, “Automatic license plate recognition,” IEEE Transactions on Intelligent Transportation Systems, vol. 5, no. 1, pp. 42-53, Mar. 2004.

[8]C. N. E. Anagnostopoulos, I. E. Anagnostopoulos, V. Loumos, and E. Kayafas, “A license plate-recognition algorithm for intelligent transportation system applications,” IEEE Transactions on Intelligent Transportation Systems, vol. 7, no. 3, pp. 377-392, Sep. 2006.

[9]J. F. Xu, S. F. Li, and M. S. Yu, “Car license plate extraction using color and edge information,” Proceedings of the International Conference on Machine Learning and Cybernetics, vol. 6, pp. 3904-3907, Aug. 2004.

[10]C. Arth, F. Limberger, and H. Bischof, “Real-time license plate recognition on an embedded DSP-platform,” IEEE Conference on Computer Vision and Pattern Recognition, pp. 1-8, Jun. 2007.

[11]V. Kamat and S. Ganesan, “An efficient implementation of the Hough transform for detecting vehicle license plates using DSP'S,” Proceedings of the Real-Time Technology and Applications Symposium, pp. 15-17, May. 1995.

[12]M. VinhDu, M. Duoqian, W. Ruizhi, and Z. Hongyun, “An improved method for Vietnam license plate location,” Proceedings of the International Conference on Multimedia Technology, pp. 2942-2946, Jul. 2011.

[13]何坤達，嵌入式車牌辨識系統，國立台灣科技大學電子工程所碩士論文，民國九十五年。

[14]陳翔傑，自動化車牌辨識系統設計，國立中央大學電機工程所碩士論文，民國九十四年。

[15]邱智國，車牌辨識系統的研製，國立台灣科技大學電機工程所碩士論文，民國九十六年。

[16]吳建中，於未設限環境中車牌辨識系統模擬與分析，國立台灣科技大學電機工程所碩士論文，民國九十五年。

[17]陳韋榤，應用快速廣義霍夫轉換於歪斜車牌定位與角度估算，國立台灣科技大學高分子工程系碩士論文，民國九十八年。