12 September 2017 Efficient subtle motion detection from high-speed video for sound recovery and vibration analysis using singular value decomposition-based approach
Author Affiliations +
Abstract
High-speed cameras provide full field measurement of structure motions and have been applied in nondestructive testing and noncontact structure monitoring. Recently, a phase-based method has been proposed to extract sound-induced vibrations from phase variations in videos, and this method provides insights into the study of remote sound surveillance and material analysis. An efficient singular value decomposition (SVD)-based approach is introduced to detect sound-induced subtle motions from pixel intensities in silent high-speed videos. A high-speed camera is initially applied to capture a video of the vibrating objects stimulated by sound fluctuations. Then, subimages collected from a small region on the captured video are reshaped into vectors and reconstructed to form a matrix. Orthonormal image bases (OIBs) are obtained from the SVD of the matrix; available vibration signal can then be obtained by projecting subsequent subimages onto specific OIBs. A simulation test is initiated to validate the effectiveness and efficiency of the proposed method. Two experiments are conducted to demonstrate the potential applications in sound recovery and material analysis. Results show that the proposed method efficiently detects subtle motions from the video.
© 2017 Society of Photo-Optical Instrumentation Engineers (SPIE)
Dashan Zhang, Dashan Zhang, Jie Guo, Jie Guo, Yi Jin, Yi Jin, Chang’an Zhu, Chang’an Zhu, } "Efficient subtle motion detection from high-speed video for sound recovery and vibration analysis using singular value decomposition-based approach," Optical Engineering 56(9), 094105 (12 September 2017). https://doi.org/10.1117/1.OE.56.9.094105 . Submission: Received: 7 May 2017; Accepted: 18 August 2017
Received: 7 May 2017; Accepted: 18 August 2017; Published: 12 September 2017
JOURNAL ARTICLE
10 PAGES


SHARE
Back to Top