为了满足物体振动的测量要求,基于双目立体视觉测量的原理,设计了一套可以在线采集、离线处理的非接触式低频振动测量的视觉系统。通过滤波去除噪声,采用Canny算子和最小二乘法椭圆拟合提取到标志点的轮廓和中心坐标,在已知特征点的空间分布规律的条件下,通过相应的坐标排序方式,对左右视图对应特征点进行准确的匹配,获得了特征点的三维坐标数据。最后对比每一帧图像中对应特征点三维坐标数据,得到待测物的振动信息。通过直线导轨作为振动源带动待测物运动的方式进行了测量实验,实验结果显示,该系统的检测精度达到±0.15 mm/m以上,基本满足低频振动测量的稳定可靠、精度高等要求。

In order to meet the requirement of vibration measurement,an online collection,offline processing of non-contact visual system of low frequency vibration measuring was designed based on the principle of binocular stereo vision.The filter was used for removing the noise and the Canny operator and the ellipse fitting based on the least squares were used for extracting the both of the contour and center coordinates of the feature points.Furthermore,according to the matching algorithm based on the epipolar constraint and the spatial distribution law of feature points,the 3D coordinate data of the feature points were obtained.Finally,the vibration information of the object was obtained by comparing the 3D coordinate data of the feature points in the each frame image.As a linear guide of the vibration source,an experiment to measure the low frequency vibration shows that the measurement accuracy can reach ±0.15 mm/m.And it can satisfy the precision and efficiency demands of the low frequency vibration measuring.