Real-world terrain model reconstruction
Rui Li · 2023
In this paper, we present a pipeline for reconstructing terrain models using LiDAR point cloud data, which aims to preserve landform features and produce realistic models. First, a normal estimation method leverages the information from the LiDAR scans in order to get more accurately oriented point clouds. Then, we employ Poisson reconstruction to generate a watertight terrain model based on the oriented point clouds. Compared to terrain models that reconstruct from Digital Elevation Models (DEMs), our proposed method can achieve a higher level of realism in the generated models.