论文部分内容阅读
提出了针对机载激光雷达点云的建筑物三维模型提取方法。首先通过滤波从点云中提取数字地形模型(DTM),再从数字表面模型(DSM)中剔除DTM影响,得到正规化数字表面模型(nDSM);然后通过高程滤波和双边滤波从nDSM中得到建筑物区域;在建筑物几何形状约定下,将建筑物分“层”处理,通过边缘探测与规格化算法得到每“层”建筑物的边缘;边缘点坐标构成了建筑物三维模型的基础。该方法借助数学形态学易于计算机处理的优点,能较为快速、准确地提取建筑物三维模型,适合大规模、快速、对精度要求不高的lidar数据提取。最后通过数据验证了该方法的可行性及有效性。
A three-dimensional model extraction method for building point cloud of airborne Lidar is proposed. Firstly, the digital terrain model (DTM) is extracted from the point cloud by filtering and the DTM effect is removed from the digital surface model (DSM) to obtain a normalized digital surface model (nDSM); then the building is obtained from the nDSM through elevation filtering and bilateral filtering In the agreement of building geometry, the building is divided into layers and the edges of each building are obtained by the edge detection and normalization algorithm. The coordinates of the edge points form the three-dimensional model of the building Foundation. With the help of the advantages of mathematical morphology, which is easy to process by computer, this method can extract 3D model of buildings quickly and accurately, which is suitable for large-scale, fast and low-precision lidar data extraction. Finally, the data verify the feasibility and effectiveness of the method.