Image-based geometric digital twinning for stone masonry elements

We present an image-based pipeline for generating geometrical digital twins (GDTs) of stone masonry elements with detail down to the level. For this purpose, we acquire RGB images individual stones and wall during construction phase. In our framework, use structure from motion (SfM) first generate 3D source destination models, which are then registered form GDT through non-linear least squares 2D point feature correspondences detected on SfM images. This method contrasts traditional techniques that register clouds using descriptors. Because robustness image descriptors, found instead features facilitates automation generation. To benchmark algorithm, compared results Euclidean–distance-based proposed metric a known textured model were synthetically generated. show feasibility full size elements, wherein GDTs generated dry-stone stone-mortar systems. study allows researchers produce accurate representations geometry walls built experimental research, reducing therefore uncertainties related size, shape arrangement minimum when comparing numerical simulations these results. Codes data sets publicly available (https://github.com/eesd-epfl/stone_masonry_GDT https://doi.org/10.5281/zenodo.7266587).