Phased-Based Video Vibration Analysis Research

Structural health monitoring is of great significance in ensuring structural safety,avoiding property damage and safety accidents.Video signal processing is an important method for non-contact measurement.The image signal can be used to obtain a wide range of vibration signals and avoid load effect.Phase-based video vibration analysis technology is a new vibration measurement method that can amplify and visualize vibrations that are not easily perceived by the human eye.It has significant results in modal analysis and vibration measurement.This thesis studies phase-based video motion processing technology.First,this thesis describes the hazards of structural damage,analyzes a variety of non-contact measurement methods,and illustrates how to process the acquired signals.Subsequently,the traditional Euler linear motion processing technology,phase-based video motion processing technology and visual microphone technology are introduced.While signals are in nonstationary and strong noise conditions,phased-based video vibration analysis method is difficult to extract target signals and will magnify noise in the frequency range.This thesis proposes a video vibration analysis method based on parameterized time-frequency analysis and principal component analysis.Phase variation signals are projected onto a small number of principal components and filtered on time-frequency domain.Not only do reconstructed signals have good time-frequency characteristics,but also computational complexity of the algorithm is greatly reduced.The experimental results show that the proposed method has significant advantages in processing nonstationary and strong noise signals.Finally,this thesis studies the problem of modal separation in phase-based video vibration analysis technology.An unsupervised modal separation algorithm based on variational modal decomposition and independent component analysis is proposed in combination with complex pyramids.The variational modal decomposition method is used to extract the main frequency components of the signals,and reduce the noise signals,then each component signal is extracted by the independent component analysis.It verifies the algorithm can separate out all modes in cantilever experiment.