基于机器视觉技术的振动台试验变形位移测量方法研究

Measuring Method Research of Deformation Displacement in Shaking Table Test Based on Machine Vision Technology

在振动台试验中,结构的变形位移是试验关注的重要数据之一;将机器视觉技术运用于振动台试验的变形位移测量上,是对传感器测量法的一个补充,从另外一种渠道力争为振动台试验提供更多的数据来源;该方法首先给定振动台一个激励信号,同时对固定在振动台台面上、粘贴有人工标志的弹性结构试件进行视频采集;然后将视频分帧成一系列静态图片,对静态图片进行图像灰度化、图像增强等图像预处理并对人工标志进行识别;接下来,在识别出所有人工标志的基础上采用最小二乘椭圆拟合法进行目标定位并进行像素标定;目标点的像素位移结合像素标定系数,最终完成变形位移的测量;最后通过振动台的固定位移试验和正弦信号激励试验证明了该方法在振动台试验中的可行性。

The deformation displacement of the structure in shaking table test is one of the important measurement parameters.Using the machine vision technology to the measurement of the deformation displacement in shaking table test is a complement of sensor measure-ment,providing more data sources from another channel.This method firstly gives vibration table an excitation signal,and collects the vid-eo for the elastic structure specimen with the artificial marks fixed on the vibration table.Secondly,to do image preprocessing such as image gray,the image enhancement about this video image sequences ,and the artificial marks recognition.Next,the least-squares ellipse fit-ting method is adopted to complete object localization on the basis of identify all artificial marks and the pixel calibration.Then,the deform-ation displacement will be calculated by the pixel displacement and the target pixel calibration coefficient.Finally,the feasibility is proved through the experiment when the fixed displacement signal and sine signal are applied to the shaking table.