机载激光扫描技术(LiDAR)可直接获取数字表面模型,但要获得高精度的数字地面模型(DTM)需要滤波去除地物点,这是一项十分复杂和耗时的工作。本文提出一种基于二次多项式的局部最小二乘拟合滤波算法,对数据进行规则格网化处理,对每个格网的数据进行局部最小二乘拟合滤波处理,并缩小格网大小迭代进行滤波。实验结果表明提出的方法能得到很好的效果。 Airborne laser scanning technology (LiDAR) can directly acquire digital surface models, but obtaining a high-precision digital terrain model (DTM) requires filtering to remove features, a complex and time-consuming task. In this paper, we propose a local least squares fitting filtering algorithm based on quadratic polynomials. The data are regularly gridded, the data of each grid is processed by local least squares fitting and filtering, and the grid size iteration is reduced Filtering. The experimental results show that the proposed method can get good results.