Cable vibration and girder displacement measurement via morphology recognition and feature matching methods in bridge structure.

Abstract

The operational safety of bridges plays an important role in national or regional economic development. Due to complex variation loads, many engineering structures experience degradation and damage with increased operating time. Therefore, it is necessary to provide effective support for the maintenance of engineering structures. Take the traffic structure as an example, a large number of bridges in cities still rely on periodic inspections. Due to limited cost, existing periodic maintenance focused on observing surface damage and paying less attention to dynamic response in operation. Therefore, a measurement system based on a smartphone for full-field deformation of the bridge was introduced. Firstly, many vision-based methods for multiple bridge members measurement were proposed and the accuracy of methods was verified through laboratory and field experiments. With the combination of template matching and feature detection methods, the displacements of the deck were measured with an error of less than 0.3%. The dynamic response can be obtained with an error of less than 1% with the mapping model. Moreover, the stability and accuracy of the proposed vision-based methods under various illumination conditions were further explored. Finally, the recommended system has significant advantages of automated processing, good portability and low cost, providing a new solution for the maintenance of the bridges in the city.