嵌入式多媒體開發平台於即時車牌辨識系統之設計與實作

近年來國人擁有自用小客車的人數日益漸增，衍生出違規停車、失竊、事故、停車場管理等社會問題，自動車牌辨識可以用來輔助處理這些問題。數位訊號處理器具有執行速度快、高效率、體積小以及適合獨立運作等優點，因此本研究採用德州儀器DM6437 EVM為開發平台，並搭配攝影機建構一套即時車牌辨識系統。實作車牌辨識系統之流程主要有三個階段：(1)車牌定位；(2)車牌字元分割；(3)車牌字元辨識。首先在影像前處理方面利用直方等化、二值化、邊緣檢測、形態學、連通物件等方法，解決所擷取的影像過亮和過暗的問題，並去除非車牌區域之背景及雜訊，再透過投影法從大小不同的車牌上，將字元分割出來，最後以簡單快速的模板匹配法找出與模板相似度最高之字元作為辨識結果。為驗證本文所提方法之可行性，我們以停車場為例進行實測，實驗結果顯示，平均辨識速度為180 毫秒，車牌定位成功率為90.5%，字元辨識率為96.5%，整體車牌辨識準確度達81.25%。

關鍵字

數位訊號處理器；嵌入式系統；影像處理；車牌辨識

並列摘要

In recent years, the number of car owners has increased, leading to a rise in social issues such as parking violations, thefts, accidents, and parking management. However, automatic license plate recognition can be employed to address these issues. Digital signal processors offer the advantages of high processing speeds, high efficiency, a compact size, and independent operation. Therefore, for this study, we selected Texas Instrument DM6437 EVM as the development platform, which we combined with a video camera to develop a real-time license plate recognition system. Implementation of the license plate recognition system comprises the following three stages: (1) license plate detection; (2) license plate character segmentation; and (3) license plate character recognition. First, image processing methods, such as histogram equalization, binarization, edge detection, mathematical morphology, and connected component labeling, were used to ensure that the captured images were neither excessively bright nor excessively dark and to remove the background and noise signals in the license plate area. Then, the projection method was used to segment the characters on license plates of varying size. Finally, template matching, which is simple and rapid, was employed to select the characters that were the most similar to the template as the recognition result. A parking lot was selected to test the feasibility of this study. The results show that the average recognition speed was 180 milliseconds, the license plate detection success rate was 90.5%, the character recognition rate was 96.5%, and the overall accuracy of the license plate recognition reached 81.25%.

並列關鍵字

Digital signal processor ； Embedded System ； Image processing ； License plate recognition

參考文獻

[4] R. C. Gonzalez, and R. E. Woods, Digital Image Processing. Reading, MA: Addison-Wesley, 1993.

[5] H. L. Bai and C. P. Liu, “A hybrid license plate extraction method based on edge statistics and morphology,” in Proc. 17th Int. Conf. on Pattern Recognition, Cambridge, United Kingdom, 2004, pp. 831-834.

[6] N. Bellas, S. M. Chai, M. Dwyer and D. Linzmeier, “FPGA implementation of a license plate recognition SoC using automatically generated streaming accelerators,” in Proc. 20th Int. Parallel & Distributed Processing Symposium, Nice, France, 2006, pp. 8-15.

[8] D. Zheng, Y. Zhao and J. Wang, “An efficient method of license plate location,” Pattern Recognition Letters, vol. 26, no. 15, Nov. 2005, pp. 2431-2438.

[9] S. Z. Wang and H. J. Lee, “Detection and recognition of license plate characters with different appearances,” in Proc. Intelligent Transportation, vol. 2, 2003, pp. 979-984.

國際替代計量

嵌入式多媒體開發平台於即時車牌辨識系統之設計與實作

未授權

主題瀏覽