基于高速视觉的非接触式振动特性识别方法

Non-contact Vibration Characteristic Recognition Method Based on High-speed Vision

针对传统视觉振动测量方法测量高频振动误差较大的问题,提出了采用高速相机搭建非接触测量系统的方法。首先,对高速相机采集图像进行预处理,降低噪声干扰;其次,通过ORB特征点检测方法检测出图像中的特征点,使用Harris评价准则代替原算法中的非极大值抑制方法,从中挑选本文算法提取出的特征点,并通过稀疏光流法对特征点的振动进行追踪,得到准确的特征点位移信息;再次,利用快速傅里叶变换计算出各特征点的振动频率;最后,以音叉和钢条为对象进行实验。实验结果表明,提出的方法在测量结果上与理论计算结果相近,最大误差仅为0.75%,运行速度较传统算法提高约33.22%。

To address the problem of large errors in measuring high-frequency vibration by traditional visual vibration measurement methods,a method of building a non-contact measurement system using a high-speed camera is proposed.Firstly,the high-speed camera acquisition image is pre-processed to reduce noise interference;secondly,the feature points in the image are detected by the ORB feature point detection method,and the Harris evaluation criterion is used to replace the non-maximum suppression method in the original algorithm,from which the feature points extracted by the algorithm are selected,and the vibration of the feature points are tracked by the sparse optical flow method to obtain accurate feature point displacement information;The vibration frequency of each feature point is then calculated using the fast Fourier transform;finally,experiments are conducted with tuning forks and steel bars,and the experimental results show that the proposed method is similar to the theoretical calculation results in terms of measurement results,with a maximum error of only 0.75%and an increase of about 33.22%in operation speed compared with the traditional algorithm.