扰动测量环境中基于图像序列的模态参数识别

针对结构模态参数识别,提出了一种基于二维连续图像序列的识别方法.该方法以CCD相机采集的图像序列为基础,在待识别结构上选取一定数量的特征点,运用因子分解法将结构运动参数和外形信息分离开,得到各特征点的运动轨迹,通过实施特征系统实现算法(ERA,Eigen-system Realization Algorithm)识别出结构的模态参数.特别是当测量仪器,即CCD相机受到扰动时,能够从图像序列中分离扰动,得到结构自身的运动参数,保证较好的识别效果.以相机受扰动时,悬臂梁模型为例进行了仿真分析.将识别出的悬臂梁前5阶频率与ANSYS计算结果进行对比,结果吻合很好,对于工程中利用光学图像识别结构动态特性具有较高的参考价值.

Modal parameter identification based on image streams under disturbance measurement environment

To aim at structure modal parameter identification,a method based on a sequence of 2D images was proposed. Chose certain feature points from the object whose modal parameter need to be identified. Factorization method was employed on the image streams taken by CCD camera to solve structure from motion problem and obtain motion track of each point. Eigen-system realization algorithm ( ERA) was used to identify modal parameter of structure. Especially when the measure equipment, that is CCD camera was disturbed, it can separate disturbance through image streams and get structure motion parameter, so that the identification result was ensured. A simulation model of cantilever beam under disturbed measurement environment was analyzed. It shows that the first five frequencies of the beam coordinating well with the ANSYS calculation result, which offers high significance to the dynamic characteristic identification using optic images in project programs.