Comparing Metrics for Evaluating 3D Reconstruction Quality in Large-Scale Environment Mapping - Archive ouverte HAL Accéder directement au contenu
Pré-Publication, Document De Travail Année : 2023

Comparing Metrics for Evaluating 3D Reconstruction Quality in Large-Scale Environment Mapping

Marianne Clausel
Cedric Pradalier

Résumé

In this study, we focus on addressing the challenge of measuring the quality of 3D reconstruction in large scale environments. Specifically, we consider scenarios where the map is built using a robot's 3D-Lidar point cloud observations, with potential uncertainty in the robot localization. When considering a large scale environment that is unstructured, such as a park or a forest, another difficulty arises: the data becomes extremely sparse. As a result, measuring the quality of a reconstruction becomes even more challenging. This study aims to compare the effectiveness of various metrics in measuring the quality of 3D reconstruction. Firstly, we evaluate these metrics in a controlled experimental setup, where the reconstruction is created by progressively degrading the ground-truth using different degradation models. Secondly, we compare their ability to measure reconstruction quality at a local level, across various simulated environments, ranging from structured to unstructured. Finally, we conduct a qualitative comparison to demonstrate the robustness of certain metrics to noise in the robot localization. Ultimately, we synthesize the properties of these metrics and provide practical recommendations for their selection.
Fichier principal
Vignette du fichier
RAS_Metrics_HAL.pdf (10.77 Mo) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04128242 , version 1 (14-06-2023)
hal-04128242 , version 2 (08-01-2024)

Identifiants

  • HAL Id : hal-04128242 , version 1

Citer

Stéphanie Aravecchia, Marianne Clausel, Cedric Pradalier. Comparing Metrics for Evaluating 3D Reconstruction Quality in Large-Scale Environment Mapping. 2023. ⟨hal-04128242v1⟩
91 Consultations
194 Téléchargements

Partager

Gmail Facebook X LinkedIn More