首页|基于角点校准的暗光下桥梁位移监测研究

基于角点校准的暗光下桥梁位移监测研究

扫码查看
基于机器视觉的数字图像方法监测结构物变形具有非接触性等优点,但受到暗光环境及人工标志物变形等因素的影响导致其应用受限.在现有模板匹配测量基础上,提出了暗光环境下基于Harris角点检测校准和Retinex暗光增强网络的桥梁位移监测方法和系统,以提高桥梁结构在复杂环境下的位移监测精度.进一步通过试验验证,结果显示10 m距离的暗光环境下桥梁位移监测最大绝对误差小于0.3 mm,在靶标偏转角度在0.6°至1.2°情况下,基于角点校准后的平均精度相比未校准提高了44%,说明该系统具备在暗光环境下进行工程结构位移监测的实用性.
Research on bridge displacement monitoring under dark light based on corner calibration
The digital image method based on machine vision to monitor the deformation of structures has many advantages such as non-contact,but its application is limited by factors such as dark light environment and artificial marker deformation.Based on the existing template matching measurement,a bridge deforma-tion monitoring method and system based on Harris corner point detection calibration and Retinex dark light enhancement network in dark light environment is proposed to improve the displacement monitoring accura-cy of bridge structures in complex environments.Further,through experimental validation,the results show that the maximum absolute error of bridge displacement monitoring under dark light environment at 10m distance is less than 0.3mm,and the mean accuracy after calibration based on angle point is improved by 44%compared to uncalibrated at the target deflection angle of 0.6 to 1.2 degrees,which indicates that the system has the practicality of engineering structure displacement monitoring under dark light environment.

bridge monitoringmachine visiondark light enhancementtemplate matchingcorner detec-tion

赵少杰、黄滏、张潇婷、马嘉成

展开 >

湘潭大学土木工程学院,湖南湘潭 411105

桥梁监测 机器视觉 暗光增强 模板匹配 角点检测

湖南省教育厅科研项目湖南省自然科学基金面上项目

19C17652021JJ30681

2024

湘潭大学学报(自然科学版)
湘潭大学

湘潭大学学报(自然科学版)

CSTPCD
影响因子:0.403
ISSN:2096-644X
年,卷(期):2024.46(2)
  • 19