基于机器视觉的低频振动台动态性能参数测试方法

Machine vision-based method for testing dynamic performance parameters of low-frequency vibration shakers

振动台为低频振动传感器校准提供必要的输入激励,其动态性能参数对校准能力与精度具有重大影响。目前,振动台动态性能参数的测量方法存在系统复杂或精度有限的问题,难以满足日益增长的高性能振动台动态性能参数测量需求。因此,提出一种基于机器视觉的低频振动台动态性能参数测试方法,通过结合高精度的摄像机标定与序列图像亚像素级动态特征提取准确地复现振动台不同频率的位移,实现振动台幅频特性、重复性、失真度的可靠解算。通过与基于传感器方法在0.01~10 Hz范围内进行对比试验,所测量的幅频特性平均误差为0.611%,重复性最大误差为0.174%,而其失真度最大误差不超过1%,表明机器视觉法在低频振动台动态性能参数测试中具有出色的表现。

The vibration shaker provides essential input excitation for the calibration of low-frequency vibration sensors,with its dynamic performance parameters significantly impacting the calibration capability and accuracy.Current methods for measuring the dynamic performance parameters of vibration shakers either involve complex systems or offer limited precision,failing to satisfy the increasing demand for highperformance measurements.In response,this study introduces a machine vision-based approach for assessing the dynamic performance parameters of low-frequency vibration shakers.This approach involves combining high-precision camera calibration with sub-pixel dynamic feature extraction from sequential images to accurately replicate the shaker's displacement at various frequencies.It facilitates reliable computation of the shaker's amplitude-frequency characteristics,repeatability,and distortion degree.Comparative tests within the0.01 Hz to 10 Hz range against sensor-based methods have shown that the measured amplitude-frequency characteristics of the vibration shaker have an average error of 0.611%,a maximum repeatability error of0.174%,and a maximum distortion error not exceeding 1%.These findings underscore the exemplary performance of the machine vision method in evaluating the dynamic performance parameters of lowfrequency vibration shakers.