Characterizing Footbridge Response from Cyclist Crossings with Computer Vision-Based Monitoring

Computer vision applications are frequently selected for short-term monitoring of bridges. The availability high resolution and frame rate consumer-grade cameras such as action makes the computer vision-based an attractive affordable option. This paper presents findings from deformations a steel girder footbridge subjected to cyclist loads captured with camera. A modified GoPro camera zoom lens was located 35 m centre bridge. High (720 × 1080 pixel) (240 fps) videos were recorded during crossings. Image processing, measurement denoising, vertical deflection interpretation influence line (bridge signature) derivations presented discussed. Both static dynamic responses identifiable (discernible in measurements) even tiny (a fraction millimetre) deflections can be accurately computed collected