Characterization and Mitigation of Speckle Noise in Laser Doppler Vibrometer on Moving Platforms (LDVom)

Abstract

Laser Doppler vibrometer (LDV) is a vibration-detecting instrument for noncontact and non-destructive measurement. It is superior to classic contact transducers in terms of the wide frequency range and high measurement resolution. LDV on moving platforms (LDVom) is one of the LDV measurement technology to one-way scan the vibrating surface, so that it is applicable in large-scale measurement like railway tracks. Speckle noise is a significant signal issue for LDV technologies, especially for LDVom. It distorts the local vibration signal dramatically and reduces the overall signal-to-noise ratio to a quite low level. The one-way scanning nature of LDVom makes it impossible to simply average the signals for noise removal. In view of the speckle noise issue of LDVom, the goal of this dissertation is to acquire new understanding of the problem and proposed there upon-based de-speckling solutions. Three aspects are investigated to achieve the research goal: 1) numerical simulation of speckle noise and characterization of noise behaviors. It can provide insight into behaviour changes of speckle noise in response to some variables and possible tools for minimizing noise strength; 2) the theoretical Fourier spectrum of speckle noise series. The resulted frequency domain characteristics can help design the de-noise signal filter accordingly; 3) development of classic approach-based and newly designed de-speckling algorithms...