ABSTRACT
Stay cables are essential load-carrying components in cable-stayed bridges. Various techniques have been developed to accurately measure the dynamic response of cables in order to monitor and assess structural health conditions in time. In this paper, an image-based integrated application is proposed for cable dynamic response measurement as well as modal identification, which provides a user friendly and cost effective solution to monitor the structural performance of cables. Compared to conventional sensors that are contact-type and pointwise, the non-contact vision-based technique, which uses a single commercial-grade camera, is able to take measurements from a remote distance while having flexible sensing capability. As a vibrating cable is the object of interest in video image processing, the detection of cable geometric features becomes the first crucial step in cable displacement extraction. A comparative study of five different feature detectors is therefore conducted to select the most suitable algorithm for detecting the vibrating cable. The feature-based video image processing technology is integrated with cable dynamic analysis for identifying modal properties of a monitored cable. The non-contact vision-based measurement technique is validated through conducting free vibration tests of a single undamped cable under laboratory conditions. Experimental results have shown that the image-based integrated application can provide reliable measurement of cable dynamic response and accurate identification of cable modal properties. As the video image sequence is the only required input, the proposed application has great potential for practical implementations in structural health monitoring.
