An advanced vision-based deformation measurement method and application on a long-span cable-stayed bridge
Shanshan Yu, Jian Zhang, Xiaoyuan He
Abstract
Abstract An advanced vision-based method focusing on fast and accurate deformation measurement was developed. First, computing speed and sub-pixel matching accuracy were improved by modifying the feature extraction process and optimizing the elimination mechanism of random sample consensus. An impact vibration test on a beam was designed to verify this improvement. Second, a new displacement transformation equation that considers the location change of measuring points for oblique optical axis measurement was proposed, with its advantage verified through a specially designed translation test. Based on the above, software for synchronous multi-region tracing was developed and applied to monitor the mid-span deflection and multi-cable tensile force of the Su Tong Bridge, the second longest cable-stayed bridge in the world. The test results show the great application prospects of the proposed optical method for remote and non-target deformation measurement.