中文摘要：

针对三维建筑物屋顶和立面目标的自动提取难题,提出了一种利用倾斜影像生成的像素高度图来识别和提取三维建筑物目标的方法,根据高度图的梯度特征提取垂直立面目标,根据高度图的最大熵阈值特征提取屋顶目标。采用多角度面阵数字航空相机获取的倾斜影像数据进行实验,验证了本文方法的正确性与有效性。

英文摘要：

Aiming at the critical issues in automatic 3D building facades and roofs extraction, a preliminary 3D building object extraction algorithm from a pixel height map, derived from the oblique airborne imagery through dense matching, was therefore designed and implemented. Based on the knowledge of building structure and surface shapes, this method takes full advantage of the height gradient feature to extract the facades, while at the same time uses the maximum entropy threshold feature to extract the roof surfaces. Multi-angle digital aerial images are employed for experimental analysis; the results show that 3D building objects with complex shape can be detected and extracted reliably from pixel height maps.