傳統的港口貨櫃檢查方法是利用人力於檢查站交換各項單據，此方法不僅耗費人力與自然資源，也可能因為人為疏失造成檢查錯誤。故本研究使用影像處理技術來達到自動執行檢查貨櫃的功能，系統中的前處理已經能夠成功地擷取出完整貨櫃背面，而本研究著重在於辨識貨櫃碼，由於貨櫃碼的位置是固定的，故以設定ROI (Region of Interesting)的方式將貨櫃碼定位出來，爾後移除其他不要的雜訊，利用垂直直方圖與水平直方圖統計貨櫃碼影像的分布情況，進而切割出圖片中的貨櫃碼。因為影像解析度很低的關係，有許多種辨識方法被提出來，本研究提出透過統計二值化像素點與樣板比較的方法來辨識貨櫃碼。依據實驗結果可以證明，處於低解析度的狀態下仍能保有七成五的辨識率，而且有可能取代掉傳統的檢查程序，以本論文提出的系統自動化執行檢查貨櫃的流程。

Traditional container checking procedures are to check related information by human inspectors at checking points. Such procedures not only waste human and nature resource but also may have mistakes caused by people negligence. This study is to use image process techniques to automatically perform those container checking procedures. Previous work has successfully extracted the rear part of the container for feature matching process. In this study, we focus on the recognition of the container number. Due to fixed location of the container number, the ROI (Region of Interesting) is defined and separated from the container image. After some noise removal process, vertical histogram and horizontal histogram are considered to count distributed situation of container codes image to accurately segment the image containing container number. Since the resolution of those number images is low, several approaches are proposed to recognize the gotten numbers. This research proposes a way to recognize codes by comparing statistics of binary pixels and template. Experiments show promising results keeping 75% recognized rate in low resolution, and then it is possible to replace the traditional checking procedure with the proposed system to check containers automatically.

[1] 交通部統計查詢網(Ministry of Transportation and Communications R.O.C), http://stat.motc.gov.tw/
[2] 陳偉明，「海關閘口以影像觸發之貨櫃辨識系統」，碩士論文，國立臺灣科技大學，2012
[3] 黃國榮，「貨櫃影像特徵比對系統之研究」，碩士論文，國立臺灣科技大學，2011
[4] 陳翔傑，「自動化車牌辨識系統設計」，碩士論文，國立中央大學電機工程研究所，2005
[5] 莊佳龍，「車輛偵測與車牌辨識系統」，碩士論文，國立中正大學機電光整合工程碩士班，2005
[6] 林清波，「基於即時影像追蹤之交通監測系統」，碩士論文，國立交通大學電機與控制工程學系，2002.
[7] 曾佩婷，「在特定車道行駛之大型車偵測與追蹤」，碩士論文，國立中央大學資訊工程研究所，2005.
[8] 呂進發，「使用影像處理技術辨認貨櫃號碼」，碩士論文，國立高雄第一科技大學電腦與通訊工程系，2004.
[9] 黃敏卿，「雙攝影機運輸監控系統之貨櫃車牌與櫃碼辨識」，碩士論文，國立臺灣科技大學電機工程研究所，2013
[10] R.C. Gonzales, R.E. Woods, Digital Image Processing 3/e, Pearson Education, 2008
[11] W.S. Gao, L. Yang, X.G. Zhang, H.Z. Liu, “An Improved Sobel Edge Detection”, 3rd IEEE International Conference on Computer Science and Information Technology (ICCSIT), pp.67-71, 2010.
[12] N. Otsu, “A threshold selection method from gray level histogram,” IEEE Trans. on Systems, Man, and Cybernetics, SMC-8, pp. 62-66, 1978.
[13] H.Z. Wang, Ying Dong, “An Improved Image Segmentation Algorithm Based on Otsu Method”, International Symposium on Photoelectronic Detection and Imaging 2007: Related Technologies and Applications, L. Zhou, Ed., vol. 6625 of Proceedings of SPIE, pp. 6
[14] 鍾國亮, “影像處理與電腦視覺”, 臺灣東華書局股份有限公司, 台北, 2006
[15] G. Bradski, A. Kaehler, Learning OpenCV, O’Reilly Media, 2008.
[16] C. Kim and J.N. Hwang, “Fast and Automatic Video Object Segmentation and Tracking for Content-Based Applications,” IEEE Transactions on Circuits and System for Video Technology, 12(7), pp.597-612, 2002.
[17] R.C. Gonzales, R.E. Woods, Digital Image Processing 3/e, Pearson Education, 2008.
[18] 維基百科(Wikipedia), “Intermodal container”, http://zh.wikipedia.org/wiki/