CAJ | 학술논문

针对合成孔径雷达（synthetic aperture radar ，SAR）图像中桥梁的特征，提出一种基于高分辨率SAR图像的桥梁自动提取算法。利用改进的Otsu阈值法将图像二值化，针对面积属性，利用连通域标记法和Otsu算法思想自适应地剔除非河流区域；通过提取河流主干线与河流边界的交点初步锁定桥梁的位置与搜索区域；利用边界集合间最小距离法提取出桥梁边界集合，通过最小二乘法拟合出桥梁两侧的边界直线。对真实的SAR图像进行处理并取得了良好的提取效果。
침대합성공경뢰체（synthetic aperture radar ，SAR）도상중교량적특정，제출일충기우고분변솔SAR도상적교량자동제취산법。이용개진적Otsu역치법장도상이치화，침대면적속성，이용련통역표기법화Otsu산법사상자괄응지척제비하류구역；통과제취하류주간선여하류변계적교점초보쇄정교량적위치여수색구역；이용변계집합간최소거리법제취출교량변계집합，통과최소이승법의합출교량량측적변계직선。대진실적SAR도상진행처리병취득료량호적제취효과。
A bridge extraction method from high resolution SAR (synthetic aperture radar) images was presented .Firstly ,the improved Otsu thresh method was employed to segment the image to a binary image with the river and the background .Second-ly ,the connected domain notation method was used to mark the connected regions and to compute their area ,and then the Otsu idea was applied to the area property instead of the grayscale so as to reject the false alarm area automatically .Thirdly ,the bridge location was confirmed by extracting the river trunk and boundary .Finally ,the bridge edge was extracted using the min-distance of the river boundary ,and two straight lines of the bridge boundary were fitted using least square estimation .Experi-mental results on obtained real SAR images confirm the effectiveness of the proposed method .