Cost-effective long-range measurement of structural displacement using a two-stage method integrating super-resolution reconstruction and sub-pixel registration.

Displacement is an imperative metric for assessing structural performance. However, the accuracy of structural displacement measurement is constrained by the physical resolution of the camera, particularly in long-range monitoring scenarios. To address these constraints, a cost-effective long-range structural displacement measurement system based on a two-stage method integrating super-resolution reconstruction and sub-pixel registration (SRSR) is proposed in this study. Initially, the proposed method achieves texture detail enhancement of low-resolution images through super-resolution reconstruction, thus improving the accuracy of visual tracking. Subsequently, a continuous convolution operator based on natural texture on the structural surface is utilized to achieve accurate sub-pixel registration. Finally, the performance of the proposed methodology is verified with field tests on a scaled truss bridge and a cable-stayed bridge. The displacement measurement accuracy is significantly enhanced through the implementation of a two-stage technique based on SRSR (including steps I and II), in comparison to the utilization of the sub-pixel registration (step II only) in isolation. The cost-effective vision-based system developed in this study provides an alternative for long-range structural displacement measurement.