Detecting faults in modular bridge expansion joints (MBEJs) is a critical challenge, as their complex dynamic behavior under operational loads is difficult to characterize with conventional methods. Existing monitoring approaches often face a trade‐off between high cost, intrusive installation, and an inability to capture the distributed kinematic patterns essential for fault diagnosis. To address this gap, this paper proposes a cost‐effective, distributed vision‐based framework designed for fault diagnosis of in‐service MBEJs. The framework employs a single camera to synchronously track an array of fiducial markers, enabling noncontact, multipoint dynamic measurement, while its algorithmic pipeline synergizes image enhancement with robust detection to ensure data reliability. Field deployment on a long‐span bridge provided initial validation of its diagnostic capability by correctly classifying a known bearing failure as a structural‐level anomaly, based on its distinctive dynamic signatures: a statistically significant elevation in displacement variance and the emergence of an anomalous, structure‐borne vibrational mode at 3.577 Hz. The framework demonstrates its potential for unmanned anomaly detection, providing critical kinematic indicators that can serve as a basis for early warnings and proactive structural management.
Huang et al. (Thu,) studied this question.