Structural state estimation of earthquake-damaged building structures by using UAV photogrammetry and point cloud segmentation

Abstract

Post-earthquake rapid structural assessment has been a widely recognized and valuable inspection need, which can accelerate community recovery and contribute to seismic resilience. Nowadays, three-dimensional point cloud models can be collected quickly using unmanned aerial vehicles (UAVs) for building structures. Regarding post-event rapid inspection, timely processing and informative interpretation of field data sets may still be a significant challenge, while limited attempts have been made to automatically extract geometrical features of point cloud models for estimating full-field deformation states of seismic-damaged structures. This study proposes a point cloud-based structural component segmentation approach to realize structural inclination and residual drift estimation at both system-level and story-level. Two real-world multi-story structures are illustrated to validate the segmentation and estimation performance of the proposed method by employing wall and column inclination and full-field residual drift distribution. The results demonstrate that the developed approach can segment the target components and present high-precision deformation estimates.