在对目前机载LiDAR数据后处理典型算法分析比较的基础上,针对机载LiDAR技术实际应用中对数据处理稳定性、智能性与适应性的更高需求,提出了一种快速稳健滤波方法,即平面延展方法。它利用单元平面的生长形成整个地表,相对常规方法增加了栅格化数据、填充空白数据及剔除错误冗余数据等操作,利用可靠的判定条件选择了基础点。采用平面延展方法对不同地形区域的LiDAR数据进行了滤波,将滤波前后的数据图像进行对比,采用三种客观误差指标评价处理质量。结果表明,平面延展方法稳健、可靠、效率高,能够在保证精度的情况下有效分离出非地面信息,在应用中有一定的优越性。 Based on the analysis and comparison of typical algorithms for airborne LiDAR data post-processing, a fast and robust filtering method is proposed for the higher demand of data processing stability, intelligence and adaptability in airborne LiDAR technology. That is the plane extension method. It makes use of the cell plane growth to form the whole surface. Compared with the conventional method, it rasterizes the data, fills in the blank data and eliminates the wrong redundant data. The base point is selected by the reliable judgment conditions. The plane extension method was used to filter the LiDAR data in different terrain areas. The data before and after filtering were compared. Three objective error indexes were used to evaluate the processing quality. The results show that the planar extension method is robust, reliable and efficient. It can effectively separate non-ground information with accuracy and has some advantages in application.