Multi-target tiny vibration video amplification method based on multi-passband filter

Abstract

The invention discloses a multi-target micro-vibration video amplification method based on a multi-passband filter, and the method comprises the following steps: S1, photographing a multi-target vibration video through a camera, carrying out the YIQ color space conversion of the video, and extracting the video data of a Y channel; S2, in the Y video, manually selecting an ROI (region of interest)region containing the edge of the first vibration target, extracting brightness signals at N pixel positions in the region, and calculating an average frequency spectrum of the signals; similarly, obtaining the average frequency spectrum of other M-1 vibration targets; and then, detecting vibration frequencies and frequency peak values of the M targets to obtain a frequency vector and a frequencypeak value vector; S3, designing a multi-passband filter according to the sum of the frequency vectors; S4, replacing a band-pass filter in the PVM method with a filter to realize amplification of theM target micro-vibration videos. The multi-target micro-vibration video amplification method provided by the invention has the advantages of realizing multi-target micro-vibration visual display andthe like.

Description

technical field [0001] The invention relates to the technical field of computer vision, in particular to a multi-target micro-vibration video amplification method based on a multi-passband filter in a real scene. Background technique [0002] There are many tiny vibrations that cannot be perceived by the human eye in nature, and these tiny vibrations often contain important information reflecting the vibration characteristics or essential properties of objects. [0003] The video amplification method is mainly used to visualize the tiny vibrations that are imperceptible to the human eye, so that the human eye can directly observe the tiny vibration phenomenon. According to the two perspectives used to describe fluid motion in fluid mechanics, existing video amplification methods are mainly divided into two categories, namely Lagrangian perspective video scaling methods and Euler perspective video scaling methods. The Lagrangian video amplification method first selects featu...

AI Technical Summary

Problems solved by technology

[0005] Although the PVM video amplification method in the Euler perspective can clearly amplify the tiny vibrations that the human eye cannot perceive, the algorithm can only amplify a tiny vibration in the video, because the band-pass filter used by the algorithm has only one passband , it can only filter out signals in the passband, and cannot screen out small vibrations in multiple passbands at the same time, so the PVM method cannot amplify multiple small vibrations at the same time

Images