第五章 模擬分析與實驗
5.4 歪斜車牌轉正執行結果
為了驗證車牌是否能正確轉正，使後續的結果能正確執行，我們這邊繪出車 牌轉正結果。目前發現若車牌本身旋轉程度不超過 50 度，且經系統截取出為車 牌者，均可以正確轉正為正車牌。之所以需要不超過 50 度，乃是因為我們假設 車牌左上角對應到正車牌的左上角，車牌右下角對應到正車牌的右下角，共四個 點，以此類推。若車牌傾斜超過 50 度，則此種對應關係可能被破壞。此外，以 實際應用上，較少出現傾斜超過 50 度的車牌出現於監控攝影機上，因此轉正的 處理結果尚稱合理。一個車牌轉正的結果如圖 5-2 所示。
(a) (b)
圖 5-2 車牌轉正結果：(a) 原圖；(b) 轉正後影像。
5.5 平行化的車牌偵測與字元辨識系統實行結果
我們對於平行化的車牌偵測與字元辨識系統實行結果如下：首先我們將一張 含有汽車與其車牌的影像輸入，而由我們所開發的平行處理系統，分別找出車牌 的位置，將歪斜的車牌拉正，將車牌裡的每一個可能字元抓出，並且使用樣板匹 配法辨識出其字元。目前結果如圖 5-3 所示，能夠順利在未經使用者任何提示的 狀況下，自動將影像中的車牌找出，並且拉正，找出其中的字元，並且辨識之。
(a) (b)
(c) (d)
圖 5-3 車牌偵測與字元辨識結果：(a) (b) 全部辨識成功； (c) 因為邊界超出切割區域 的外圍而部分未受到辨識； (d) 因為樣板匹配問題而部分辨識錯誤。
其中，圖 5-3 (a) 中的所有影像字元全部辨識成功，在圖 5-3 (b) 中，雖然輸
入的車牌呈現歪斜的狀況，但由於我們提出的系統會先將車牌拉正，因此系統也 能正確辨識出車牌中的文字。而由於有時會因為抓取邊界不足的關係，在圖 5-3 (c) 中會漏找邊緣字元，但成功擷取的字元皆辨識正確，此種問題可望藉由手動 增加框選車牌範圍來解決之；在圖 5-3 (d) 中有部分字元辨識錯誤的狀況，可能 的原因是因為樣板匹配之問題，此種問題比較常出現在新舊車牌中，字元外觀不 相同之情況。若能針對同一種字元，增加樣板數，亦可望能提升辨識之準確率。
第六章 結論與未來展望
本論文提出一連串的系統與方法，利用影像顯示卡及 CUDA 的平行處理機 制，針對時下需要藉由平行處理的方式來加速的影像處理及影像辨識系統，有 效實現出加速的效果，不僅縮短相較於僅使用 CPU 時執行時間的數倍，執行 效果亦不因此有所減損。
依據實驗結果，本文得到以下結論，我們將全景攝影系統中全向圖與全景圖 轉換的演算法使用 CUDA 的程式架構與 8 個串流處理器進行運算時間是 CPU 的 1/6.33，將車牌辨識系統中我們透過圖形處理器加速車牌文字辨識的計算速度相 較於 CPU，我們可以得到 100 倍的改善。透過平行化的車牌辨識系統能夠減少電 腦判斷時間並降低在動態識別中的誤判，透過平行化後的全向圖轉換演算法能夠 在全景攝影系統上解決高畫質的全向圖影像傳換成全景圖影像 CPU 無法即時轉 換的問題，希望未來本研究論文在消費性電子產品上能夠有所應用。
參考文獻
[1] K. Steffen, E. Markus, H. Sorin A, and S. Integrierte Schaltungen, " Real time image processing based on reconfigurable hardware acceleration," in Proceedings of Workshop Heterogeneous Reconfigurable Systems on Chip, 2002.
[2] P. Cuong , A. Zaid and B. Koen "Heterogeneous hardware accelerator architecture for streaming image processing," in Proceedings of International Conference on Advanced Technologies for Communications, pp. 374-379, 2013.
[3] C. Edwards, "Graphical future," in Engineering & Technology, vol. 4, no. 10, pp. 339, 6-19 June 6 2009.
[4] S. K. Nayar, "Catadioptric omnidirectional camera," in Computer Vision and Pattern Recognition, 1997. in Proceedings of Conference on IEEE Computer Society, 1997, pp. 482-488.
[5] V. Nalwa, "A true omnidirectional viewer," technical report, Bell Laboratories1996.
[6] S. Baker and S. K. Nayar, "A theory of catadioptric image formation," in Proceedings of Sixth International Conference on Computer Vision, pp. 35-42, 1998.
[7] A. Krishnan and N. Ahuja, "Panoramic image acquisition," in Computer Vision and Pattern Recognition, 1996. Proceedings CVPR'96, 1996 IEEE Computer Society Conference on, 1996, pp. 379-384.
[8] A. Goshtasby and W. A. Gruver, "Design of a single-lens stereo camera system," Pattern Recognition, vol. 26, pp. 923-937, 1993.
[9] H. F. Ho and W. J. Chou, "Method to enhance panoramic photography resolution and vertical field of view," in Proceedings of the 27th Conference on Computer Vision, Graphics, and Image Processing, Pingtung, Taiwan , 2014.
[10] Y. Yagi and S. Kawato, "Panorama scene analysis with conic projection," in Intelligent Robots and Systems' 90.'Towards a New Frontier of Applications', Proceedings. IROS'90. IEEE International Workshop on, 1990, pp. 181-187.
[11] P.-H. Huang, "The Research and Development of Panoramic Total Internal Reflection Lens and Omnidirectional Imager System," 2011.
[12] G. L. Mariottini, E. Alunno, and D. Prattichizzo, "The epipolar geometry toolbox (EGT) for Matlab," Citeseer2004.
[13] C. Saravanan, "Color Image to Grayscale Image Conversion," in Proceedings of Second International Conference on Computer Engineering and Applications, Bali Island, pp. 196-199,
2010.
[14] Johndcook.com. (2009). Converting color to grayscale. [online] Available at:
http://www.johndcook.com/blog/2009/08/24/algorithms-convert-color-grayscale/ [Accessed 2 Apr. 2016].
[15] H. Yeganeh, A. Ziaei and A. Rezaie, "A novel approach for contrast enhancement based on Histogram Equalization," Computer and Communication Engineering, 2008. ICCCE 2008.
International Conference on, Kuala Lumpur, 2008, pp. 256-260.
[16] (finalfrank), f. (2016). [OpenCV] 直方圖等化 @ Frank's 資訊科技潮流站 :: 痞客 邦 PIXNET ::. [online] Frank's 資訊科技潮流站. Available at:
http://finalfrank.pixnet.net/blog/post/30668256-%5Bopencv%5D-%E7%9B%B4%E6%96%B9%E5%9C%96%E7%AD%89%E5%8C%96 [Accessed 15 Jun.
2016].
[17] (finalfrank), f. (2016). [影像處理] OTSU演算法程式 @ Frank's 資訊科技潮流站 ::
痞 客 邦 PIXNET ::. [online] Frank's 資 訊 科 技 潮 流 站 . Available at:
http://finalfrank.pixnet.net/blog/post/29574285-
%5B%E5%BD%B1%E5%83%8F%E8%99%95%E7%90%86%5D-otsu%E6%BC%94%E7%AE%97%E6%B3%95%E7%A8%8B%E5%BC%8F [Accessed 22 Jun. 2016].
[18] N. Otsu"A Threshold Selection Method from Gray-Level Histograms," in IEEE Transactions on Systems, Man, and Cybernetics, vol. 9, no. 1, pp. 62-66, Jan. 1979.
[19] Fahn, Chin-Shyurng, and Meng-Luen Wu. "An Autonomous Aesthetics-driven Photographing Instructor with Personality Prediction." The 2013 International Conference on Computer Graphics, Visualization, Computer Vision, and Game Technology. Atlantis Press, 2013.
[20] S. Suzuki, K. Abe, "Topological structural analysis of digitized binary images by border following," Computer Vision, Graphics, and Image Processing, vol. 30, no. 1. pp. 32-46, 1985.
[21] Fernando, S. (2016). Shape Detection & Tracking using Contours. [online] Opencv-srf.blogspot.tw. Available at: http://opencv-srf.blogspot.tw/2011/09/object-detection-tracking-using-contours.html [Accessed 25 Jun. 2016].
[22] P. Viola, M. Jones, “Robust Real-Time Face Detection. International Journal of Computer Vision”, 57(2), 2004, 137-154.
[23] J. Gao, J. Yang, “An Adaptive Algorithm for Text Detection from Natural Scenes”, Proceedings of the 2001 IEEE Computer Society Conference on Computer Vision and Pattern
Recognition, 2001.
[24] V. Shapiro, G. Gluhchev, D. Dimov, “Towards a Multinational Car License Plate Recognition System”, Machine Vision and Application, 17, 2006, 173-183.
[25] K. Nadeem AM and R. J and H. Hans A, " License plate recognition system," in Proceedings of SPIE's International Symposium on Optical Science, Engineering, and Instrumentation pp. 14-24, 1998.
[26] E. Patrick and J. Carlos and B. Stefan and M. Donal, " A Hotelling transformation approach for rapid inversion of atmospheric spectra," Journal of Quantitative Spectroscopy and Radiative Transfer, vol. 73, no. 6, pp. 529-543, 2002.
[27] C. Paolo, F. Paolo, G. Michael N and S. Flavio, " Optical recognition of motor vehicle license plates," in IEEE Transactions on Vehicular Technology, vol. 44, no. 4, pp. 790-799, 1995.
[28] H. Yo-Ping, L. Shi-Yong and C. Wei-Po, "A template-based model for license plate recognition," in Proceedings of International Conference on Networking, Sensing and Control, vol. 2, pp. 737-742, 2004.
[29] Upload.wikimedia.org. (2016). [online] Available at:
https://upload.wikimedia.org/wikipedia/commons/2/2a/Svm_max_sep_hyperplane_with_mar gin.png [Accessed 3 Mar. 2016].
[30] Upload.wikimedia.org. (2016). [online] Available at:
https://upload.wikimedia.org/wikipedia/commons/thumb/6/67/5_Stage_Pipeline.svg/300px-5_Stage_Pipeline.svg.png [Accessed 5 Jul. 2016].
自 傳
我
是周文瑞我來自高雄， 2013 年就讀國立臺灣師範大學電機工程學系碩士 班，主要研究將全景攝影系統與車牌辨識系統的平行計算，由於本身很喜歡寫程 式所以平時最大的興趣就是自己一個人在房間裡開發軟體，很感謝林政宏教授能 夠在碩士班擔 任我 的指導教授， 讓我 能夠學習 到 CUDA 以及平行計算 的相 關專業知識，希望畢業後將所學貢獻社會。學 術 成 就
(1.論文發表、2.專利、3.競賽、4.參與研究計畫、5.其他特殊表現)
Cheng-Hung Lin and Wen-Jui Chou, "Acceleration of the Transformation from Elliptic Omnidirectional Images to Panoramic Images Using Graphic Processing Units", in Proc. of IEEE International Conference on Consumer Electronics - TAIWAN (IEEE 2016 ICCE-TW), NANTOU COUNTY, TAIWAN, May 27-29, 2016.