2D/3D data fusion for the comparative analysis of the vaults of Notre-Dame de Paris before and after the fire.

• Methodology for a complete geometrical analysis of the vaults of Notre-Dame de Paris after the fire. • 2D image-based method for comparison before and after the fire. • 3D point cloud-based method for comparison before and after the fire. • New method for accurate alignment of very large point clouds. The work presented in this article was carried out following the fire (April 15, 2019) that ravaged Notre-Dame de Paris. The goal was to establish in a short and limited time-span the structural diagnosis of the remaining vaults of the cathedral. The analysis relied on 2D and 3D data acquired before and after the fire, and available at that time. Two complementary methods have been developed to estimate potential displacements from multi-temporal surveys. The first method is mainly based on the processing of image collections obtained with photogrammetric routines right after the fire. The second method, based on the processing of raw 3D data acquired with Terrestrial Laser Scanning, was developed afterward to refine and complete the analysis. Both methods have benefited from essential functionalities offered by two open-source solutions proven to enable "real-based" modeling in Cultural Heritage contexts, namely Micmac and CloudCompare. This paper presents the complementary use of these two software packages whilst detailing some procedural computing approaches to document and support structural diagnosis in an emergency situation. The limits and capabilities of each method are exposed and analyzed, toward the objective to develop 2.5D and 3D based geometrical analysis, respectively from photogrammetric and lasergrammetric sources. The results are presented and discussed regarding their potential usage for a quantitative and accurate diagnosis of vaults. [ABSTRACT FROM AUTHOR]