Optimal positioning of terrestrial LiDAR scanner stations in complex 3D environments with a multiobjective optimization method based on GPU simulations
Résumé
Currently, the scanning of complex industrial sites is commonly performed using terrestrial LiDAR scanners. As the quality of the resulting point cloud depends mainly on the number and positions of LiDAR stations, this scanning process can be preliminarily optimized by means of a 3D model. A previous study proposed multiobjective optimization based on the linear scalarization of three functions to maximize coverage and overlapping of point cloud stations while minimizing their number. Because these objectives conflict, this study proposes the use of MO-CMA-ES, a global multiobjective optimization algorithm, to provide a full Pareto front and allow the user to make an informed decision. Our method is the first to rely on realistic LiDAR simulations that operate in fully 3D complex environments and provide point clouds with optionally noisy coordinates. For performance considerations, ray-traced simulations and objective evaluations were performed using a GPU. Furthermore, clash detection in the proximity of station positions was also considered. After validating our method's behavior and demonstrating its superiority over the conventional approach in a simple case, we conducted a study on an industrial-grade case based on a 2.7-million-triangle model, further demonstrating our method's effectiveness by producing a minimal 15-station solution with optimal coverage and overlapping.
Origine : Fichiers produits par l'(les) auteur(s)