摘要
针对传统的桥梁振动测试方法可达性差、效率低的问题,提出了一种基于高帧视频分析的桥梁动力特性非接触识别算法.通过制定相应的连通域算法实现动态定位与跟踪,精准识别安装在结构表面具有特殊几何形状的标志点,获取结构振动时程信息,并在多种光照环境下进行实验室简支梁锤击振动试验以及户外斜拉人行桥行人激励振动试验.结果表明:对于简支梁试验,条件良好时所提算法得到的基频与实测值的误差小于0.2%,条件不佳时识别误差仍可小于1.1%;对于人行桥试验,所提算法得到的第1阶自振频率小于1.2%.说明所提算法基频数据识别准确,实施效率高,经济性好且鲁棒性强,可实现恶劣环境下可达性较差桥梁的非接触检测.
Aiming at the technical problems of poor accessibility and low efficiency of traditional bridge vibration testing methods,a non-contact recognition algorithm for dynamic characteristics of bridges based on high-frame video analysis was proposed.The corresponding connected domain algorithm was formulated to realize the dynamic positioning and tracking of the structure.The landmarks with special geometric shapes installed on the structure surface were accurately identified and the time history information of the structure vibration was obtained.The vibration test of a simple supported beam in the laboratory and the pedestrian vibration test of an outdoor cable-stayed pedestrian bridge were carried out in various lighting environments.The results show that,for the simply supported beam test,the test error between the fundamental frequency obtained by the proposed algorithm and the measured value is less than 0.2%when the experimental conditions are good,and the recognition error can still be less than 1.1%in inadequate experimental conditions.For the pedestrian bridge test,the first-order natural frequency obtained by the proposed algorithm is less than 1.2%.The proposed algorithm can accurately identify the fundamental frequency data.It has advantages of high efficiency,good economy and robustness during implementation,thus achieving non-contact detection of bridges with poor accessibility in harsh environments.
作者
熊文
程瑜
Xiong Wen;Cheng Yu(School of Transportation, Southeast University, Nanjing 211189, China)
出处
《东南大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2020年第3期433-439,共7页
Journal of Southeast University：Natural Science Edition
基金
国家自然科学基金资助项目(51978160)
浙江省公路科技计划资助项目(2018H10).
关键词
非接触
机器视觉
高帧视频
图像处理
桥梁振动
模态识别
non-contact
machine vision
high-frame video
image processing
bridge vibration
modal analysis
作者简介
熊文(1982—),男,博士,副教授,wxiong@seu.edu.cn.
