A structure knowledge-synthetic aperture radar interferometry integration method for high-precision deformation monitoring and risk identification of sea-crossing bridges

Abstract

Deformation monitoring and risk identification of sea-crossing bridges are essential to mitigate hazards and prevent loss of human life and property. Satellite-based Synthetic Aperture Radar Interferometry (InSAR) technology can detect millimeter-scale deformation, showing unique advantages in the safety monitoring of sea-crossing bridges. However, the existing InSAR methods only extract point-like targets (PTs) based on the coherent index, but ignores the analysis of multiple SAR incoherent information and the foreground-background scattering characteristics differences of bridges, leading to low-density and low-accuracy of PTs on sea-crossing bridges. Moreover, most InSAR-based studies identified structural risks according to deformation measurements without fully considering the various safe deformation ranges of different structural components, resulting in high false-alarm/miss-detection rates in structural risk identification of sea-crossing bridges. To address these issues, a structure knowledge-InSAR integration approach is developed for high-precision deformation monitoring and reliable risk identification of sea-crossing bridges. Firstly, the SAR incoherent information and foreground-background scattering characteristics of the bridge structure are analyzed and applied to improve the density of extractable PTs and remove the incorrect noise signals. Then, the bridge structural mechanics model is combined with the InSAR time-series displacements to analyze the mechanical property degradation of different bridge components, improving the reliability of InSAR-based structural risk identification. This approach is applied to the Stonecutters Bridge and Tsing Ma Bridge using the TerraSAR-X and COSMO-SkyMed images from 2011 to 2012 and the Sentinel-1A images from 2015 to 2017. The results indicate that the densities of PTs extracted on the two bridges increased by about 40% using the new approach, and incorrect noise signals are removed. Moreover, the mechanical properties of different bridge components can be evaluated through the analysis of their structural stress and time-series displacements, helping to decrease the false-alarm/miss-detection rates of InSAR-based structural risk identification. The bridge deformation is correlated with the temperature variation when the temperature difference is large (≥10 °C), but no longer dominated by thermal dilation when the temperature difference is less than 10 °C due to the influence of environmental effects.