欧拉运动放大算法在桥梁振动分析中的试验研究

Experimental Study on Euler Motion Amplification Algorithm in Bridge Vibration Analysis

为发展新的桥梁动态位移识别和分析方法,解决桥梁非接触式动态位移测量中的振动显微问题,将欧拉运动放大算法应用于桥梁振动分析。首先利用普通摄像机连续采集一系列桥梁数字影像。而后通过欧拉运动放大算法,对桥梁的连续图像序列进行空域分解,使用对微弱运动敏感的时域带通滤波器进行滤波,得到连续图像序列中的微小振动,再对低频的振动信号采用信号放大处理。随后,通过对振动放大后的数字影像进行帧数分解,获得常规视频采样频率下的桥梁试件振动时间历程序列,最后通过对图像序列中标记点的边缘特征分析,得到了试验桥梁标记点的动态位移测量结果。通过与位移传感器测量结果和百分表的比较,证明所采用和提出的分析方法是可行的,能够满足桥梁动态位移测量的应用需求,也能够针对桥梁的某些振动特性开展初步的定量分析,做到对桥梁结构的微小振动位移进行识别。

In order to develop a new method for identifying and analyzingdynamic displacement of bridges and solve the micro-vibration problem innon-contact dynamic displacement measurement of bridges,Euler motionmagnification algorithm is applied to bridge vibration analysis. First,aseries of digital images of bridge structures are collected continuouslyby ordinary camera. Then,by using Euler motion magnificationalgorithm,the continuous image sequence of bridge is decomposed inspatial domain,the time-domain bandpass filter sensitive to weak motionis used for filtering to obtain the micro-vibration in the continuousimage sequence,and then the low-frequency vibration signal is amplified.Subsequently,the time history image sequence of the bridge specimenunder the conventional video sampling frequency is obtained by framedecomposing the amplified digital image. Finally,the dynamicdisplacement measurement result of the labeled points on the test bridgeis obtained by analyzing the edge characteristics of the labeled pointsin the image sequence. Compared with the measurement results ofdisplacement sensor and the dial indicator,it is proved that thepresented analysis method is feasible. It can meet the applicationrequirements of bridge dynamic displacement measurement,it can alsocarry out preliminary quantitative analysis of some vibrationcharacteristics of bridges and to identify the micro vibrationdisplacements of bridge structures.