Real-Time Vision-Based Deformation Measurement of Long-Span Bridge with Inertial Sensors

Objectives: The deformation measurement of long-span bridge alignment is of great significance to the assessment of bridge health and safety. The traditional deformation measurement methods of long-span bridge alignment, such as sensor and close range vision measurement, have shortcomings such as high measurement cost and limited measurement frequency, which are difficult to meet the requirements of high-precision measurement of deformation of long-span bridge alignment. Methods: Combined with dynamic precision engineering measurement and machine vision technology, this paper proposes a method for measuring the deformation of long-span bridge alignment based on inertial cameras. The camera is used to take high-frequency images of bridge measurement targets. At the same time, the inertial sensors are used to correct the measurement errors caused by camera movement. The deformation of long-span bridge alignment is calculated through inertial cameras in series mode, and the corresponding measuring equipment is developed and applied to the health monitoring system of a super large bridge. Results: The precision experiment results show that root mean square error of the proposed method is 0.38 mm. In engineering applications, the deformation of bridge alignment trend measured by the inertial camera is highly consistent with the static leveling results. The correlation coefficient is better than 99.90%, and the mutual difference is 4.94 mm. Conclusions: The proposed method can measure the linear deformation of large span bridges with high accuracy and real-time, and has good engineering value. © 2023 Wuhan University. All rights reserved.