Data fusion in structural health monitoring: Integrating computer vision and strain gauges for enhanced inspection

In the evolving field of Structural Health Monitoring (SHM), the integration of diverse sensing techniques is pivotal for comprehensive structural analysis. This study introduces a novel data fusion methodology that synergizes the precision of strain gauges with the broad coverage of computer vision techniques. Aimed at advancing SHM practices, our approach facilitates a detailed understanding of structural behavior by concurrently analyzing strain fields and displacement signals captured through video streams. We employ Physics Informed Neural Networks (PINNs) to refine measurements and ensure physical plausibility in our data interpretation. Our methodology’s effectiveness is validated through laboratory experiments on a simplified structural model, demonstrating enhanced accuracy and reliability in SHM. This paper highlights the potential of integrating traditional and contemporary sensing techniques in infrastructure monitoring, setting a new benchmark in the field.