一种光学遥感影像建筑区域自动提取方法  

An Automatic Built-up Area Extraction in Optical Satellite Images

作　　者：蔡利平[1] 苗则朗[2] 史文中[2,3] 柳思聪[1] 张华[1] 

机构地区：[1]中国矿业大学环境与测绘学院,江苏徐州221116 [2]香港理工大学土地测量与资讯学系,香港九龙999077 [3]武汉大学香港理工大学空间信息联合实验室,湖北武汉430072

出　　处：《地理与地理信息科学》2015年第4期22-26,F0003,共6页Geography and Geo-Information Science

基　　金：国家自然科学基金重点项目(41331175);江苏高校优势学科建设工程资助项目(SZBF2011-6-B35);江苏省普通高校研究生科研创新计划项目(CXZZ13_0939)

摘　　要：提出一种利用光学遥感影像自动获取建筑区域的方法。该方法利用基于结构张量的和算子、乘算子预计建筑区域的概率,按照融合规则将两种概率图融合,通过主动轮廓从融合的概率图中提取建筑区域。选取多幅ALOS影像对此算法进行试验分析,并与基于Gabor滤波法和Harris角点检测法进行比较。试验结果表明,与其他两种方法相比,该方法从光学遥感影像上自动提取建筑区域的结果具有更高精度。This study presents an automatic approach to extract built-up areas from optical satelhte images. The proposea mem- od consists of three main steps. First, the built-up area probability is estimated using the sum operator and multiply operator based on structural tensor. Second, the fusion rule is introduced to integrate two probability maps. Finally, built-up areas are ex- tracted from the fusion result by the use of active contour. The performance of the proposed method is tested using multiple ALOS satellite images and compared with Gabor filtering based method and Harris corner detection based method. The experi- mental results demonstrate that the proposed method performance is superior to the other two methods in terms of extraction accuracy.

关 键 词：主动轮廓 建筑区域提取 融合 光学遥感影像 结构张量 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]