Drone-Based Precision Deflection Measurement for Bridges

Aging civil infrastructure, especially bridges, poses significant challenges to safety and economic stability. In the United States and Japan, over 40% of bridges are more than 50 years old. In Europe, 10% of bridges are over 100 years old. Structural deficiencies, which affect 7.5% of US bridges, require urgent maintenance. To ensure infrastructure sustainability, effective, non­destructive evaluation methods are essential for assessing structural health.¹ Traditional sensing devices face limitations due to costs and constraints. Vision-based methods have advanced, but they still rely on stationary cameras.² Unmanned aerial vehicles offer a promising alternative for displacement measurement, overcoming these limitations with enhanced maneuverability.³ However, achieving sub-millimeter accuracy in drone-based aerial photography remains a significant challenge.

Recently, we presented a novel framework for achieving high-precision displacement measurements using drone video footage.⁴ By attaching moiré markers to a bridge and using the sampling moiré (SM) method,⁵ the study attained precision levels of up to 1/100 pixels. An active balancing-compensation strategy was developed to effectively differentiate actual displacement from camera motion to further enhance measurement accuracy. Comprehensive analyses and experimental validations underscore the system’s reliability, establishing it as a viable and cost-effective solution for autonomous bridge inspections.

Specifically, this optical imaging technique introduces a novel technology for high-precision image de-blurring, inspired by the human inner ear’s balancing system, to enhance drone-based bridge displacement measurements. The method employs two 2D reference markers, analogous to the ear’s vestibular receptors, to stabilize images during drone hovering by simplifying six degrees of freedom (DoF) into four. Markers placed on bridge abutments emulate the ear’s equilibrium function, correcting drone-induced errors in 3-axis translations and rotation, ensuring accurate deflection measurements.

In the stabilization process, the center coordinates of the reference markers are first detected with pixel-level accuracy. A similarity transform is then applied, followed by the SM method, refining the stabilization to 1/100-pixel accuracy. The experimental setup involved drone displacement measurement on a 30-meter-long steel bridge in Japan. We used a 6K camera to capture video as a 20-ton test vehicle crossed the bridge. The method successfully measured a maximum deflection of 3.2 mm after stabilization, highlighting its potential for practical bridge inspection.

We believe our research demonstrates the feasibility of millimeter-scale displacement measurement using drone photography. The proposed method, which incorporates phase information and a four DoF parameterization scheme, achieves high accuracy, low computational complexity and robustness to noise. This technology has significant potential to improve the regular and safe inspection of aging infrastructure, addressing the challenges posed by deteriorating bridges worldwide.

Researchers

Shien Ri, Jiaxing Ye and Nobuyuki Toyama, National Institute of Advanced Industrial Science and Technology, Japan

Norihiko Ogura, CORE Institute of Technology Corporation, Japan

References

1. Y. Jeong et al. J. Struct. Integrity Maint. 3, 126 (2018).

2. S. Ri et al. Strain 56, e12351 (2020).

3. S. Laflamme et al. Meas. Sci. Technol. 34, 093001 (2023).

4. S. Ri et al. Nat. Commun. 15, 395 (2024).

5. S. Ri et al. Exp. Mech. 50, 501 (2010).