Fast and robust vision-based cable force monitoring method free from environmental disturbances

Abstract

For long-term vision-based cable force monitoring, it is essential to fully consider different types of challenges, such as complex backgrounds in the field of view, illumination changes, occlusion, camera motion, and real-time realization. To address these challenges, this study proposes a vision-based, robust real-time cable force measurement method. In terms of innovation, three unique aspects were incorporated. First, to solve the difficulty of cable edge recognition caused by complex backgrounds, a linear feature detection operator called edge drawing lines was optimized by setting screening strategy based on the slope and line segments length. Second, the average of bilateral edge displacements was used as the final value to offset the displacement deviation caused by the illumination change. Third, to overcome obstacles to identifying the cable frequency caused by the camera motion, an interference frequency elimination method without reference targets or additional sensors was proposed. Furthermore, the integrated cable force calculation method was developed into a real-time measurement software. Finally, the proposed method was applied to a long-span cable-stayed bridge, where the camera was vibrated by the bridge deck or the wind significantly. The reliability of the proposed method was proved by comparing the results obtained with those achieved by accelerometers.