Computer Vision-based Structural Vibration Measurement And Modal Identification

In recent years,non-contact structural vibration measurement method based on computer vision has become a hotspot in the field of structural health monitoring because of its convenient and efficient characteristics.The traditional vibration monitoring methods have relatively high cost and limited space measurement points.The non-contact vibration measurement based on computer vision is relatively inexpensive and has high accuracy.It can realize remote full-field measurement and plays an increasingly important role in practical engineering.Therefore,this thesis studies the method of structural vibration testing and modal parameter identification based on computer vision.The main research contents of this paper include:The basic theory of video motion amplification technology is studied,including Euler image amplification method,video motion amplification technology based on phase,and video amplification technology of Reisz pyramid.The basic principle and implementation steps of video amplification are expounded.Video data from time domain to obtain,the algorithm based on multi-scale decomposition data in frequency domain,the frequency domain decomposition for unused space,get the video objects in subtle changes in order to obtain important information structure vibration,and amplification of the information,make the video image contains important invisible structure vibration information is obvious.This algorithm can display in real time the small changes on a specific frequency scale selected by multiple scales.The theory of structural micro-vibration measurement based on computer vision is proposed.The method based on the theory of Euler method of image magnification principle,originally by Euler method of image magnification for amplification of subtle changes in frequency domain scales is extracted,and the tiny structure vibration information,this information includes the basic dynamic characteristics of structure,thus provides the structure modal parameter identification.Using NExT+ERA method,the structural modal parameters(frequency,damping ratio and mode shapes)are identified based on the vibration displacements measured by computer vision method.The structural micro-vibration information extracted and modal identification from the video image are achieved.