Thermal response measurement of bridges using vision-based monitoring

Bridges are usually subjected to complex load scenarios in their lifetime due to the combined effect of traffic load and environmental loads such as temperature and wind. Temperature related thermal load is a major factor influencing the long-term behaviour of bridges, while traffic load influences their immediate response. This research investigates the feasibility of measuring the thermal response of bridges such as displacement and strain with the vision-based monitoring (VBM) technology using simulated data. VBM is a non-contact approach for measuring structural response using computer vision. A typical VBM system consists of hardware components for image collection and image processing algorithms for the calculation of structural responses. VBM has shown promising accuracy in measuring the static and dynamic response of bridges, however, its accuracy in thermal response measurement is not extensively investigated. Thermal response at nodal and element level is calculated using simulated data of a railway steel truss bridge for a period of one year. Different VBM scenarios with a varying field of view (FOV) are created to measure the simulated responses by changing the camera properties and camera-to-target distance. The accuracy for thermal response measurement using 4K resolution cameras has been calculated at 1/100 pixel resolution for the considered FOVs. The results show that the required accuracy is more than 1/100 pixel for an HD resolution camera when the camera-to-target distance exceeds 17.4m and 54.4m for 18mm lens and 55mm lens respectively.