Non-contact intelligent detection technology for railway arch bridge performance degradation based on UAV Image recognition

Bridges are crucial components of high-speed railway projects, and their structural integrity significantly impacts the operational safety of high-speed railways. This paper introduces a non-contact intelligent detection technology for assessing the deterioration of high-speed railway bridges using unmanned aerial vehicle (UAV) image recognition. The methodology involves collecting image data using a UAV and digital camera and processing them technically to generate consistent point-cloud data. Subsequently, these data are integrated into a unified point-cloud model through point-cloud alignment. Finally, a refined three-dimensional (3D) model of a high-speed railway bridge was developed by fusing heterogeneous data through live 3D reconstruction. The method has the advantages of high detection speed and fewer personnel requirements; this technology can be used for daily monitoring of the technical basis and can arrange a small number of personnel to complete the daily inspection. The empirical results demonstrate that this inspection method is not constrained by skylight points and provides a real-time and highly efficient reflection of the conditions of the bridge. The recognition accuracy and image acquisition range satisfy the inspection requirements for the operation and maintenance of high-speed railway bridges.