Full-field motion with a high spatial resolution can reflect the health state of long-span bridges. Traditional structural health monitoring (SHM) systems measure the structural displacement at sparse points only. Despite the development of various methods for obtaining high-resolution responses, they fail to estimate the multi-scale motions of real long-span bridges. A novel full-field motion estimation method is proposed to directly measure the entire structural responses using one digital camera. In this approach, multi-scale motions of different components under unknown excitations are encoded in a complex pyramid. The parameters of the complex pyramid are determined by the localization index to record robust local phases at each measurement point. Two indices, namely, the spatial Laplacian of the phase and the amplitude value degradation index, are designed to adaptively estimate multi-scale displacements contained in the complex pyramid. The proposed method estimates the full-field displacements of a steel frame accurately in an experimental test. Moreover, the full-field displacement of the 1377-m long main span Tsing Ma Bridge is estimated using only one shot. Results suggest that the displacement estimated by the proposed method matches well with the GPS data and further visualizes the bridge mode shapes with a high resolution.

中文翻译：

---

使用一台相机和稳健的自适应复杂金字塔对大跨桥梁进行全场位移测量

具有高空间分辨率的全场运动可以反映大跨桥梁的健康状态。传统的结构健康监测（SHM）系统仅测量稀疏点的结构位移。尽管开发了各种方法来获得高分辨率响应，但它们无法估计真实大跨度桥梁的多尺度运动。提出了一种新颖的全场运动估计方法，使用一台数码相机直接测量整个结构响应。在这种方法中，未知激励下不同组件的多尺度运动被编码在复杂的金字塔中。复杂金字塔的参数由定位索引确定，以记录每个测量点的鲁棒局部相位。设计了两个指标，即相位的空间拉普拉斯和幅度值退化指标来自适应估计复杂金字塔中包含的多尺度位移。该方法在实验测试中准确地估计了钢框架的全场位移。此外，1377米长的青马大桥主跨的全场位移仅用一次拍摄进行了估算。结果表明，该方法估计的位移与 GPS 数据匹配良好，并进一步以高分辨率可视化桥梁模态形状。