A machine vision method for the measurement of vibration amplitude

Measurements of vibration amplitude are important in many industrial applications. In particular, motion blur is a class of important information that has almost been ignored in conventional methods for motion perception. In this paper, a novel machine vision method is proposed to measure vibration amplitude based on image motion blur. In order to reveal the relationship between vibration and image blur, blurred images are analysed in the spatial and frequency domains. A series of dark parallel ripples which appears on the spectrum of the vibration-blurred image is exactly related to the vibration amplitude. Using image processing techniques, information on these ripples and parameters of vibration amplitude are obtained. Both simulation and experiment results are provided in detail to demonstrate the feasibility and efficiency of the proposed method.