机载激光雷达(Lidar)点云数据滤波是Lidar数据后处理研究的重点和难点之一,也是首要解决的问题。传统曲面约束滤波算法利用最小二乘法拟合地形曲面,易受种子点粗差影响。针对这一问题,引入抗差估计理论改善曲面拟合效果,并设计自适应阈值确定的方法区分地面点与地物点。使用国际摄影测量与遥感学会(ISPRS)测试数据进行实验,与传统的8种经典滤波方法进行对比,实验结果表明,抗差估计能得到更为合理的拟合曲面,获取的滤波结果非常可靠,对各种地形的适应性较强,具备较高实用价值。 Airborne Lidar point cloud data filtering is one of the key and difficult points in the research of Lidar data postprocessing. It is also the primary problem to be solved. The traditional surface constrained filtering algorithm uses the least square method to fit the terrain surface and is easily affected by the gross error of the seed points. In order to solve this problem, the theory of robustness estimation is introduced to improve the fitting effect of surface, and the method of adaptive threshold determination is designed to distinguish between ground point and feature point. Experiments with ISPRS test data are compared with the traditional eight classical filtering methods. The experimental results show that the proposed method can get a more reasonable fitting surface and the obtained filtering results are very reliable. Adaptable to a variety of terrain, with high practical value.