A Laser-Based High Precision Monitoring System for Railway Subgrade Settlement Based on Machine Vision
J. Tao, Y-Z. Min and W-Z Ren
To improve the measurement accuracy of subgrade surface settlement deformation monitoring, this study proposes a high precision, long-term, automatic online monitoring system for subgrade surface settlement imaging. First, the working principle and working mode of the online monitoring system for subgrade surface settlement are introduced. Second, the key technologies of the monitoring system is introduced. For the problem that the relative position of the target-illumination camera in the monitoring system is deflected via train vibration, the pose measurement system is proposed and a method of position and attitude calculation (p4p) for rectangular four feature points of the target surface is proposed. Finally, the experimental results indicate that the pose measurement method proposed in this study can solve the rotation angle error less than 0.09°, and the relative angle error is 1.003% when the relative deflection angle is less than 2°, which can be effectively applied to the subgrade images in the settlement online monitoring system. The in situ testing results illustrated that the monitoring system corrected by the pose measurement system avoids the detection of non-real subgrade settlement and improves the measurement accuracy. We also compared the cumulative error of the monitoring system with the manual observation error for 10 months and found the error values were less than 1 mm, which meets the requirements of subgrade surface settlement monitoring.
Keywords: Diode laser, railway, subgrade settlement, machine vision, pose estimation, monitoring system