低能见度前下视航拍图像中机场跑道的自动检测.pdfVIP

33 3 ( ) Vol. 33 No . 3 2005 3 J. Huazhong Univ. of Sci. Tech. ( Nature Science Edit ion) Mar. 2005 1 1 2 梁超圣 柳 健 王志成 ( 1 华中科技大学 电子与信息工程系, 湖北武汉430074; 2 华中科技大学 图像识别与人工智能研究所, 湖北武汉430074) : . , Otsu , ; , , , ; , , . : . : ; ; ; ; Otsu ; ; : TP751 : A :2005) Automati c detecti on of ai rf ield runways in forwardlooking aerial images wi th l ow visibility L iang Chaosheng Liu Jian Wang Zhicheng Abstract: To solve the roblem of detecting the airfield runways automatically in forw ardlooking aerial im ages with low visibility, a method w as ro osed. The algorithms can be cast into three ste s. The recursive segmentation method based on Otsu w as used to extract the main outline of airfield image from com lex clutters. Mathematical mor hology methods such as dilation and o ening o erations were used to deal w ith the airfield image segmented by recursive Otsu method. O ening o eration can remove some isolated noise w hile dilation o eration can connect the airfield com onents into a w hole. Region growth method was used to track and locate the rimary and secondary runways of airfield. The ex eriments show that the algo rithms have good erformance of detecting and locating airfield runways in forw ardlooking aerial images w ith low visibility . Key words: forwardlooking aerial images; low visibility; automatic detection; airfield runw ays; Otsu seg mentation; mor hological filter; region grow ing Liang Chaosheng Postgraduate; De t. of Electronics Information Eng . , Huazhong U niv. of Sci. Tech. , Wuhan 430074, China. [ 1] . , , , , , . [ 2] , ,