Machine Vision-Based Deformation Monitoring and Validation for Ultra-High Pylons of the Zhangjinggao Yangtze River Bridge

Advances in Civil Engineering

卷:2026

Research Article

Xinpeng You¹  Xiuping Huang¹  Hang Yuan¹  Tao Wang¹  Dapeng Zhu² 

[1] CCCC Second Harbour Engineering Co., Ltd., Wuhan, China

[2] GNSS Research Center, Wuhan University, Wuhan, China, whu.edu.cn

关键词: cable-supported bridges;    construction safety;    deformation monitoring;    intelligent construction;    machine vision;    structural health monitoring;    ultra-high pylons;    Zhangjinggao Yangtze River Bridge;   

DOI  :  10.1155/adce/4846792

 received in 2025-12-17, accepted in 2026-06-19,  发布年份 2026

来源: Hindawi Publishing Corporation

【 摘 要 】

Deformation monitoring of ultra-high pylons in super-long-span cable-supported bridges faces considerable challenges due to structural complexity, stringent installation accuracy requirements, and susceptibility to environmental interference when using conventional methods. This paper presents the development and validation of a real-time deformation monitoring method for ultra-high pylons based on machine vision technology, with the Zhangjinggao Yangtze River Bridge serving as a case study. High-resolution industrial cameras form the core hardware of the system, which are equipped with specially designed retroreflective targets and integrated with environmental humidity and temperature sensors. The system implements Zhang’s camera calibration method and subpixel localization algorithms. The core contributions of this study are: (1) the first high-precision monitoring over a distance of 300 m during the active construction of a 350 m ultra-high pylon; (2) a hybrid subpixel localization algorithm (0.1-pixel precision) robust to varying illumination conditions; and (3) a complete uncertainty quantification framework incorporating atmospheric turbulence and thermal drift. Field test results demonstrate a measurement accuracy better than 2 mm for the south pylon under all tested conditions. The average measurement errors were 0.25 mm during the daytime and 0.15 mm at night. The system successfully detected cyclical deformations caused by temperature variations and captured transient responses induced by construction activities, with a peak deformation fluctuation of 3.28 mm. Quantitative agreement with theoretical models is confirmed (mean absolute difference ≤ 0.35 mm, R2 ≥ 0.94). The proposed methodology offers significant improvements in construction safety, quality control, and demonstrates broad applicability, holding promise for widespread adoption in engineering practice.

【 授权许可】

CC BY   
Copyright © 2026 Xinpeng You et al. Advances in Civil Engineering published by John Wiley & Sons Ltd. 2026

【 图 表 】

Figure 11

Figure 10

Figure 8

Figure 8

Figure 7

Figure 6

Figure 7

Figure 4

Figure 4

Figure 2

Figure 1