Study of Point Cloud Integration between UAV-based Dataset and Aerial LiDAR Survey

Abstract

The migration of Photogrammetry from the analog medium to the digital medium changed how photogrammetric surveys were carried out and how data are processed, allowing the automation of several steps used in the photogrammetric design workflow. There is extensive research in cartographic and geodetic sciences, involving the acquisition of data through photogrammetry, and through techniques such as Structure from Motion (SfM) and Laser Scanner survey. Such research and work involve the use of point clouds from Unmanned Aerial Vehicle (UAV) photogrammetric surveys and Terrestrial Laser Scanner (TLS) surveys, as well as the fusion of these point clouds. This work aims to study the integration of an aerial LiDAR point cloud with a Photogrammetric Point Cloud (PPC) from the UAV survey to minimize occlusion failures and building edges, densify the point cloud, and reduce spurious points. Statistical analyses were performed for the refinement, adjustment, and integration between the point clouds. The results showed that the correction map generated for integrating the clouds reached the objective proposed in this work. The integration of clouds increased from 7 pts/m² of the cloud coming from LiDAR to 140 pts/m² in the final.