Seismic assessment of unreinforced masonry façades from images using macroelement-based modeling

This study proposes an automated, image-based framework for rapid seismic assessment of unreinforced masonry (URM) façades. By leveraging computer vision and photogrammetry, the methodology transforms multi-view external RGB images into 3D polygonal models, applies semantic segmentation to identify material and geometric features, and discretizes façades into macroelements for nonlinear structural simulations under cyclic in-plane loading. The approach was validated with experimental data from European typologies, employing a workflow compatible with ADAPTIC and OpenSees software, and assessed for computational efficiency. The case studies focused on single-story URM façades, using parameters from the Enhanced Safety and Efficient Construction of Masonry Structures in Europe (ESECMaSE) database, and simulated wall responses to benchmark seismic vulnerabilities.

The key findings show that external imagery alone can facilitate scalable, rapid assessments of façade vulnerabilities, providing visual representations of macroelement damage and force-displacement responses. The authors highlight that while the framework predicts in-plane damage accurately and efficiently, it cannot substitute comprehensive seismic analyses due to the absence of internal structural data and limited out-of-plane failure assessment. They recommend expanding the framework to integrate additional imagery sources and internal modeling, improve material property calibration, and validate dynamic responses with real ground motion data. The study demonstrates potential for large-scale disaster resilience planning in urban areas, especially where manual data collection is impractical, and offers a basis for future developments, including real-time seismic risk mapping from urban street-view and aerial imagery.