Stay cable tension estimation using a vision-based monitoring system under complex environments

Abstract Cables are the primary load-bearing component of cable-stayed bridges, the accurate monitoring of cable tension is essential for evaluating cable service reliability and bridge safety. Addressing the challenges faced by existing vision monitoring technologies in complex environments, such as limited camera resolution, low-light conditions, adverse weather and camera vibration, a non-contact vision-based monitoring system based on computer vision technology has been developed and applied to cable tension monitoring in cable-stayed bridges. The monitoring system addresses the challenges under low-light and adverse weather by utilizing an industrial camera equipped with an infrared filter and an infrared light source. A sub-pixel template matching algorithm is employed to enhance measurement accuracy under limited resolution. Furthermore, a camera vibration cancellation algorithm is introduced to mitigate the impact of external vibrations, significantly improving measurement stability. Both model experiments and field test were conducted to validate the monitoring system. Results from the model experiments demonstrate high measurement accuracy and stability under limited resolution, low-light, and adverse weather, with a maximum frequency recognition error of only 0.4%. The field test further demonstrates a significant improvement in measurement accuracy through the camera vibration cancellation method, achieving a cable tension error of only 0.86% compared to the design tension. These results highlight the system’s potential for reliable cable tension monitoring in complex environments.