Lightweight vision-based displacement measurement system with edge computing for long-term structure health monitoring

Structural health monitoring (SHM) provides real-time data on infrastructure conditions, enabling timely and cost-effective maintenance interventions for enhanced safety and extended service life. Computer vision-based non-contact sensors have emerged as a promising alternative to conventional contact-type sensors for structural displacement measurement and SHM. However, most existing vision-based systems rely on temporary camera setups and offline post-processing, limiting their applicability to long-term real-time monitoring. This paper presents a lightweight smart vision-based displacement measurement system (LightSVDMS), which integrates a lightweight and smart monitoring device to perform video acquisition, real-time processing, and displacement identification at the edge, significantly reducing data transmission requirements. The system was validated through shaking table tests, bridge scale model experiments, and long-term field measurements over six months, enabling multi-point full-field displacement monitoring. Experimental results demonstrate a displacement measurement error of less than 1.5% in scale model tests, verifying the measurement accuracy of the proposed system, while the six-month field deployment confirms its long-term stability and practical usability in real bridge environments. The novelty of this work lies in the systematic co-design of a passive-target, edge-computing-based displacement monitoring framework that overcomes the limitations of traditional methods in real-time performance and rapid installation, offering a cost-effective, easily deployable solution for condition assessment of aging bridge infrastructure.