北京航空航天大學，中法工程師學院(理科試驗班)，100191
　　
　　　摘要：
　　近年來，隨著計算機科學與技術的迅速發展，圖像處理與機器視覺已成為相關的熱門研究領域，并受到人們越來越多的重視和關注。圖像分割作為圖像處理中的關鍵技術之一，是人們進行圖像理解、圖像分析以及進一步圖像識別的基礎。
　　在本文圖像分割方法中，基于區域合并[1]提出了新的合并條件，對于基于顏色的統計區域合并中，通過對待合并區域顏色值的差異分析，進而與建立于該理論而推出的與區域大小自適應的閾值進行比較，從而判斷是否合并。而本文中除了對于顏色特征得考慮外，還對區域對排序中梯度的算法進行了改進，引入了彩色圖像中向量梯度的算法；對于區域對的合并除了顏色閾值外，還增加了幾何分布對合并的約束，考量了對艦船這類對象的圖像進行分割的個性。
　　　關鍵詞：邊界檢測，輪廓提取，區域合并，向量梯度， 幾何特性
　　　　The segmentation of the color naval boat image based on region merging
　　　LIU Haiwei
　　Beihang university, school of sino-french engineering(Science Experimental Class), 100191, liuhaiwei215@gmail.com
　　　Abstract
　　Recently, with the rapid development of the computer science and technology, the image processing and machine vision have attracted many attentions, which are applied in a variety of corresponding fields. As one of key technologies in image processing, image segmentation is a base  of image understanding,  image analysis and image recognition.
　　In the method illustrated in this thesis, new condition for the region merging[1] is proposed. For the statistical region merging based on the color features, by calculating the difference between the pair of the regions waiting to be merged, and comparing with the adaptive threshold obtained based on the statistical theory, we can make a decision for their merge. And in this thesis, apart from the color features, new algorithm of calculating the gradient of the color image is introduced: the vector gradient; and apart from the adaptive color threshold, the constraint of the geometry features is also given into consideration for the specialization of the naval ship object.
　　　Keywords:  Edge detection, contour extraction, region merging, vector gradient, geometry features

參考文獻：

[1]Richard Nock and Frank Nielsen, Statistical Region Merging, IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, VOL. 26, NO. 11, NOVEMBER 2004
[2]Shao-qing H E S Y, Tian-wei S H I A L I. Detection of Ship Target Under Sea Background Based on Texture High-Order Fractal Feature [J][J]. Optics & Optoelectronic Technology, 2008, 4: 025.
[3] R. C. Gonzalez,R. E. Woods.數字圖像處理.阮秋琦等譯. 第二版. 北京: 電子工業出版社, 2003.
[4] N. R. Pal,S. K. Pal. A review on image segmentation techniques. Pattern Recognition, 1993, 26(9):1277-1294.
[6]Fahad A, Morris T. A faster graph-based segmentation algorithm with statistical region merge[M]//Advances in Visual Computing. Springer Berlin Heidelberg, 2006: 286-293.
[7] Krauskopf J, Williams D R, Heeley D W. Cardinal directions of color space[J]. Vision research, 1982, 22(9): 1123-1131.
[8] Chan T F, Vese L A. Active contours without edges[J]. Image Processing, IEEE Transactions on, 2001, 10(2): 266-277.

作者簡介：
劉海偉，于2007年考入北航中法工程師學院，本碩連讀，并在此期間于2010-2012年去法國馬賽中央理工交換兩年，2012年10月-現在，繼續就讀于北航中法工程師學院，信息與通信工程專業。