Retaining Wall Deformation Monitoring with Camera Array Systems

Advancements in time-series 3D geospatial data have significantly improved deformation monitoring by enabling long-term and detailed analysis of structural and ground movement. However, conventional techniques such as LiDAR and photogrammetry often face challenges like coordinate system inconsistencies across time frames and the need for post-processing to align datasets, which can compromise accuracy and efficiency. To overcome these limitations, this study introduces a camera array system employing Structure-from-Motion (SfM) technology for monitoring retaining wall deformation. The system generates high-accuracy, real-scale, and co-registered 3D point clouds, facilitating 4D (3D spatial + time) geometric modeling. This allows for direct point-to-point deformation tracking and eliminates the need for complex frame-to-frame registration, enabling precise assessment of structural changes. The system is cost-effective, scalable, and capable of near real-time operation, making it particularly useful for continuous monitoring of geotechnical hazards such as slope instability and retaining wall failures. Optimized data handling and preprocessing techniques address challenges related to large dataset management and variable site conditions, such as lighting and weather. This work presents a potential solution for enhancing geotechnical risk assessment and structural health monitoring, offering improved safety, efficiency, and decision-making capabilities for infrastructure management and civil engineering applications.