Abstract Manager

Bridge Weigh-In-Motion (BWIM) systems are critical for monitoring traffic loads and enforcing weight regulations. The accuracy of BWIM systems relies on the precise calibration of strain influence surfaces. Traditional calibration methods require controlled loading scenarios, which are costly, time-consuming, and disruptive to traffic. This paper proposes a novel, scalable methodology for calibrating the bending strain influence surface of highway bridges using computer vision under regular traffic conditions, eliminating the need for controlled testing. The method utilizes real-time traffic data and computer vision to detect and localize standard vehicles crossing the bridge. The strain responses generated by these vehicles are then used to calibrate the strain influence surface, accurately capturing the bridge's behavior under operational conditions. This approach provides a cost-effective and efficient alternative to conventional methods, ensuring scalability for deployment across a wide range of bridge types and traffic environments.

The proposed methodology is validated using a case-study bridge located on the Trans-Canada Highway in New Brunswick, Canada. The bridge is instrumented with a long-term monitoring system, including accelerometers, strain gauges, and cameras, to facilitate data collection and analysis. Results confirm that the method delivers accurate calibration, significantly enhancing the reliability and effectiveness of BWIM systems in monitoring actual traffic loads. This methodology addresses key limitations of existing BWIM calibration practices by enabling calibration under normal traffic conditions. Its robustness and scalability position it as a practical solution for improving the precision and applicability of BWIM systems, supporting efficient enforcement of weight regulations and optimizing traffic load monitoring.