作　　者： ; ; ; ; ;

机构地区： 清华大学信息科学技术学院自动化系

出　　处： 《计算机学报》 2005年第7期1199-1204,共6页

摘　　要： 高层建筑物形状的复杂多变使传统的模型匹配方法难以提取.由建筑物垂直边缘检测出发,采用证据推理法,该文提出了一种在单视角高分辨率航空影像中自动提取高层建筑物的策略.该方法充分结合物方空间和图像空间各种有用信息,综合运用信号处理、计算机视觉等领域多种先进技术,逐步推理得出建筑物的位置、高度和屋顶轮廓等信息,从而实现了建筑物的三维重构. The diversity and complexity of high-rise building shapes make it very difficult to extract them by traditional modeling and matching algorithms. Starting from detection of vertical lines, a strategy for automatic extraction of high-rise building in monocular high-resolution aerial images is developed under the evidential reasoning methodology, while many useful information in object space and image space are combined, and multiple advanced techniques of signal processing and computer vision are incorporated to reason out locations, heights and roof boundaries of buildings step by step. Sequentially, 3D reconstruction is achieved.

关 键 词： 航空影像 建筑物自动提取 垂直边缘 窗户纹理 屋顶轮廓

领　　域： [自动化与计算机技术] [自动化与计算机技术]