Abstract: In order to find a new production method for DEM and improve the updating efficiency of city DEM, this paper proposes a new method to filter non-ground elevation points from ground elevation points in DSM by progressive morphological filter. The ground elevation points are then triangulated to Delaunay triangulation and the DEM is generated by intersecting rays with the Delaunay triangulation. The experiment shows that the proposed method not only generates DEM effectively, but also generates DEM with great precision, proving that it has important practical value. In conclusion, the proposed algorithm is technically feasible and it can improve the utilization rate of drone DSM to promote the renewal of city DSM.

Key words: DSM, DEM, progressive morphological filter, grid sampling

摘要： 为了探索一种全新的DEM(Digital Elevation Model,简称DEM)生产方法,提高城市DEM的更新效率,利用渐进形态学滤波分离DSM中地面高程点和非地面高程点,然后对地面高程点构建Delaunay三角网,并基于射线与三角网求交的方式插值生成完整的DEM。实验表明,文中方法不仅能够有效地完成DEM的数据生产,且生产的DEM误差可控,能够满足实际的应用需求。因此,该方法具有极强的技术可行性,可以提高无人机DSM的利用率,促进城市DEM的动态更新。

关键词: DSM, DEM, 形态学滤波, 网格采样