基于图像矫正与特征融合的结构振动特性监测OACHSSCDCSTPCD

In order to deal with the difficulties of camera calibration in field measurement,a structural modal parameters identification method is established based on machine vision measurement.A normalized orthogonal projection model is proposed to simplify the imaging model.In addition,the multi-point local calibration instead of the overall calibration is applied to lower the diffi-culty of cameral calibration.A method is proposed to improve the stability of feature points matching,i.e.,the distortion of the imaging of artificial target is rectified,and the weighted average of the gradi-ent information and position information is used as the basis for feature points matching.The accuracy and effectiveness of the proposed method is verified by a vibration test of a three-floor frame structure.The results show that the displacement and modal parameters obtained by the machine vision method are basically consistent with the real values.In addition,the proposed method still has high accuracy and stability under different shooting angles.This method is promising in the dynamic displacement monitoring of super high-rise buildings and long-span bridges.