Bridge health monitoring is essential for ensuring structural safety, yet high cost and energy consumption continue to limit large-scale deployment. This study develops a lightweight hierarchical adaptive strain monitoring framework that integrates microscopic vision-based sensing with event-driven monitoring control. A low-cost vision-based strain sensing node is developed and validated through laboratory and field deployment, achieving high measurement accuracy, with a maximum error below 10 με in static tests and below 15 με in dynamic tests, and a mean error below 5 με. Building on this sensing capability, hierarchical adaptive monitoring strategies are implemented at both the single-node and network levels. Individual sensors adjust sampling rates according to environmental variations, while significant traffic-induced strain responses detected at key locations trigger coordinated high-frequency acquisition across the monitoring network. Field results show that the framework reduces data volume and energy consumption by more than 45% during environmentally dominated monitoring periods, while still reliably capturing transient strain responses associated with heavy vehicles and dense traffic conditions. The proposed framework enables reliable response capture together with sustainable long-term operation, and provides a practical basis for scalable and energy-efficient bridge strain monitoring under real-world conditions.
Liu et al. (Fri,) studied this question.