DEM Construction Method for Slopes Using Three-Dimensional Point Cloud Data Based on Moving Least Square Theory

Abstract

The multitemporal digital elevation model (DEM) of slopes constructed from three-dimensional (3D) point cloud data (PCD), which is obtained by terrestrial laser scanning (TLS), can realize the real-time monitoring and analysis of landslides. This paper proposes a new slope DEM construction method for the 3D ground PCD based on moving least square (MLS) theory. The proposed DEM construction method has three important components: MLS-based planar projection interpolation, self-adaptive hole-filling, and greedy projection triangulation (GPT) algorithms. On the basis of MLS theory, the planar projection interpolation algorithm can obtain interpolation-reconstructed points that maintain the important topographic features of slopes, and the self-adaptive hole-filling algorithm can identify hole-filling points from the interpolation-reconstructed 3D points and repair holes of the original 3D PCD. The modified GPT algorithm can efficiently construct more complete DEMs based on the hole-repaired 3D PCD with the support of the Delaunay criterion. The effectiveness of the proposed method is demonstrated with a case study. The DEMs constructed by the proposed method are continuous and can well simulate the topographic features of slopes. Thus, the new method can provide effective DEM data support in many fields, such as surveying and mapping, and geotechnical engineering.