本研究提出了一個使用樣板比對方式的交通標誌辨識系統，用以對台灣道路上的警告標誌--外形為紅色三角形，內部為白底黑色圖形；紅色禁制標誌--外形為紅色圓形，內部為白底黑色圖形；及藍色禁制標誌--外觀為藍色圓形，內部為藍底白色圖形等三種交通標誌進行偵測與辨識，並實現於工研院自行研發之PAC Duo嵌入式開發板上。系統之組成分為兩大子系統；第一部份為交通標誌偵測系統，其主要工作為由輸入之影像中偵測是否存在任何要檢測之交通標誌。方法是利用色彩切割方式，將影像中之紅色資訊與藍色資訊分別找出，並將這些切割下來之紅色資訊區域與藍色資訊區域經篩選的動作，例如去掉面積過大或過小的區域，以留下最具有可能是存在交通標誌的候選區域，接著將這些交通標誌的候選區域中之紅色資訊部份與警告標誌之樣板(三角形)及紅色禁制標誌之樣板(圓形)進行比對；並將候選區域中之藍色資訊部份與藍色禁制標誌之樣板(圓形)進行比對，以取得交通標誌偵測之結果。第二部份為交通標誌辨識系統，主要之工作是用以決定前述所檢測出來之交通標誌的內容，依交通標誌外型種類不同分別使用三角形樣板與圓形樣板對交通標誌內部之感興趣圖形產生特徵向量，並與資料庫中之特徵向量進行關聯性比對，以辨識被檢測出來之交通標誌的內容。交通標誌之內容是以一個具有五個分量之向量來表示之，它是將交通標誌正規化後的標誌內容區域分割成幾個小區域，並選其中對標誌內容有貢獻之五個小區域以產生五個分量，每一個分量為記錄該小區域內之像素元中佔有標誌內容之像素元的百分比。本研究中所處理之警告標誌有42種，紅色禁制標誌有32種，藍色禁制標誌有13種，對每一種交通標誌均考慮五個不同之旋轉角度，以解決用不同角度對目標物取得影像之問題。特徵向量間之比對是採用計算兩向量間之絕對誤差和來求得兩向量間之關聯性的程度。PAC Duo嵌入式系統上擁有一個ARM微處理器與兩個DSP處理器，它們之間是透過共享記憶體方式存取資料，ARM微處理器與兩個DSP處理器間之溝通是透過Mailbox方式通訊，另外，DSP之本身所擁有之記憶體(local memory)空間太小，使得處理資料之速率無法達到理想。本系統測試之動態影片的解析度為320 x 240像素元，拍攝環境有晴天與陰天，平均之辨識正確率為83%。平均處理一個畫框之時間為0.6秒，系統之處理速率有待進一步提升，才能達即時之需求。

In this research a traffic sign recognition system based on template matching was presented. The system was implemented on the ITRI’s PAC Duo embedded system for recognizing the traffic signs including the warning signs (42 kinds)-- Red triangle sharp and white background with back pattern, red prohibition signs(32 kinds)-- Red circle sharp and white background with back pattern, and blue prohibition signs(13 kinds)-- Blue circle sharp and blue background with white pattern used in Taiwan. The proposed traffic sign recognition system was divided into two subsystems. The first one is called traffic sign detection subsystem. It was used to determine whether any traffic sign existed in the source image. Color segmentation method was used to extract the red information and the blue information of the input image, respectively. For these too large and too small areas of the extracted regions, they seems not as the candidates of traffic signs, are eliminated to reserve the candidate regions of traffic signs. Then the warning sign template and the red prohibition sign template were used to find the candidate sign regions of the red information, and the blue prohibition sign template was used to find the candidate sign regions of the blue information. The second subsystem called traffic recognition subsystem. It was used to determine the content of the detected traffic signs. Each normalized sign has been represented by 5 features and 5 different angle versions are considered in the DB. The correlation between the candidate and the object signs was calculated to determine the content of the traffic sign. The ITRI’s PAC Duo consists of one ARM processor and two DSP processors. The communication mechanism of the two kinds of processors is via mailbox and with shared memory. The resolution of the test videos is 320x240 pixels per frame, and captured on sunny day and cloudy day. The experimental result shows that the average procession time is 0.6 seconds for a frame and the average recognition rate is 83%. For real time application, the performance of the embedded system must be improved.

[1] Y. Wang, L. Liu and Y. Zhao, "Traffic sign detection based on fixed color combination and intensity restraint," International Symposium on Computer Network and Multimedia Technology, Wuhan, pp. 1-5, 18-20 Jan.2009.
[2] F. Qin, B. Fang and H. Zhao, " Traffic Sign Segmentation and Recognition in Scene Images," Chinese Conference on Pattern Recognition, Chongqing, pp. 1-5, 21-23 Oct. 2010.
[3] Tam T. Le, Son T. Tran, Seichii Mita and Thuc D. Nguyen, " Real time traffic sign detection using color and shape-based features," ACIIDS'10 Proceedings of the Second international conference on Intelligent information and database systems: Part II, Hue City, Vietnam, pp. 268-278, 24-26 March.2010.
[4] X. Gao, K. Hong, P. Passmore, L. Podladchikova and D. Shaposhnikov, " Colour Vision Model-Based Approach for Segmentation of Traffic Signs," Journal on Image and Video Processing - Color in Image and Video Processing, Volume 2008, January 2008.
[5] S. Lafuente-Arroyo, P. García-Díaz, F.J. Acevedo-Rodríguez, P. Gil-Jiménez and S. Maldonado- Bascón, " Traffic sign classification invariant to rotations using support vector machines," ACIVS’04, Brussels, Belgium, August 2004.
[6] Y. Aoyagi and T. Asakura, " A study on traffic sign recognition in scene image using genetic algorithms and neural networks," in Proc. 22nd IEEE Int. Conf. Ind. Electron., vol. 3, Taipei, Taiwan, pp. 1838–1843, 5-10 Aug.1996.
[7] M. Heath, S. Sarkar, T. Sanocki, and K. Bowyer, " A robust visual method for assessing the relative performance of edge-detection algorithms," IEEE J. Pattern Anal. Mach. Intell, vol. 19, no. 12, pp. 1338–1359, 1997.
[8] L. Pacheco, J. Batlle and X. Cufi, " A New Approach to Real Time Traffic Sign Recognition Based on Colour Information," Proceedings of the Intelligent Vehicles Symposium, pp. 339-344, 24-26 Oct.1994.
[9] A. Vavilin and K. Jo, " Automatic Detection and Recognition of Traffic Signs using Geometric Structure Analysis," SICE-ICASE International Joint Conference, Busan, pp. 1451-1456, 18-21 Oct.2006.
[10] S. Lafuente-Arroyo, P. Gil-Jimenez, R. Maldonado-Bascon, F. Lopez-Ferreras and S. Maldonado-Bascon, " Traffic Sign Shape Classification Evaluation I: SVM using Distance to Borders," Proceedings of IEEE Intelligent Vehicles Symposium, pp. 557-562, 6-8 June 2005.
[11] J. Hatzidimos, " Automatic traffic sign recognition in digital images," Proceedings of the International Conference on Theory and Applications of Mathematics and Informatics, Thessaloniki, Greece, pp. 174–184, 2004.
[12] P.G. Jime´nez, S.M. Bascón, H.G. Moreno, S.L. Arroyo and F.L. Ferreras, " Traffic sign shape classification and localization based on the normalized FFT of the signature of blobs and 2D homographies ," Signal Processing, vol. 88, pp. 2943-2955, 2008.
[13] Y.B. Damavandi and K. Mohammadi, " Speed limit traffic sign detection and recognition," IEEE Conference on Cybernetics and Intelligent Systems, Singapore, pp. 797-802, 2004.
[14] C.H. Lai and C.C. Yu, " An Efficient Real-Time Traffic Sign Recognition System for Intelligent Vehicles with Smart Phones," International Conference on Technologies and Applications of Artificial Intelligence , Hsinchu City, pp. 195-202, 18-20 Nov. 2010.
[15] T.P. Cao and G. Deng, " Real-Time Vision-Based Stop Sign Detection System on FPGA," Computing: Techniques and Applications, Canberra, ACT, pp. 465-471, 2008.
[16] S.M. Bascon, S.L. Arroyo, P.G. Jimenez, H.G. Moreno, F.L. Ferreras, " Road-Sign Detection and Recognition Based on Support Vector Machines," IEEE Transactions on Intelligent Transportation Systems, vol. 8, pp. 264-278, 2007.
[17] S.M. Bascón, J.A. Rodríguez, S.L. Arroyo, A.F. Caballero, F.L. Ferreras, " An optimization on pictogram identification for the road-sign recognition task using SVMs," Computer Vision and Image Understanding, vol. 114, pp. 373-383, 2010.
[18] A.D. Escalera, L.E. Moreno, M.A. Salichs, J.M. Armingol, " Road traffic sign detection and classification," IEEE Transactions on Industrial Electronics, vol. 44, pp. 848-859, 1997.
[19] C. Bahlmann, Y. Zhu, V. Ramesh, M. Pellkofer and T. Koehler, " A system for traffic sign detection, tracking, and recognition using color, shape, and motion information," In Proceedings of the IEEE Symposium on Intelligent Vehicles, pp. 255-260, 2005.
[20] M. Sonka, V. Hlavac and R. Boyle. " An Embedded System for Real-Time Traffic Sign Recognizing," 3rd International Design and Test Workshop, Monastir, pp. 273-276, 20-22 Dec. 2008.
[21] W.Y. Wu, T.C. Hsieh and C.S. Lai, " Extracting Road Signs using the Color Information, " World Academy of Science, Engineering and Technology, Issue 32, Aug. 2007.
[22] 工研院, "Intro_PAC_Duo_platform.ppt", 2008.
[23] 工研院, " PACDUO_Quick_Start_Guide_v2.pdf", 2009.