Blind Source Separation Methods for Video-based Structural Dynamics: A Comparative Study
Thaisse Paes, Moisés Felipe Silva, Joao Weyl Costa

DOI: 10.14209/sbrt.2021.1570724169
Evento: XXXIX Simpósio Brasileiro de Telecomunicações e Processamento de Sinais (SBrT2021)
Keywords: Video-based Structural Dynamics Blind Source Separation Dimensionality Reduction Modal Analysis
Operational modal analysis has been effectively used for video-based structural dynamics identification in recent years. With several different approaches, the ones based on the blind source separation strategy have received increased attention for identifying structural dynamics characteristics. This work presents a comparison study on the blind source separation techniques for videos of vibrating structures. Two dimensionality reduction techniques are used for video compression along with six source separation algorithms, resulting in twelve different frameworks tested over a laboratory cantilever beam structure. For specific algorithms, the results indicate that dimensionality reduction techniques play a major role in the mode estimation performance.