Research On Dynamic Displacement Monitoring Method For Isolation Structures Based On Computer Vision

In recent years,with the continuous development of seismic isolation technology,seismic isolation bearings as the main energy-consuming components of seismic isolation structures have been widely used in high-rise buildings,large bridges,nuclear power plants and other important engineering fields.Nowadays,structural health monitoring has become a more effective method to extend the service life of seismic isolation bearings,mostly using traditional contact displacement sensors and acceleration sensors for monitoring.With the breakthrough development of computer vision technology,it has the advantages of non-contact,high accuracy and easy operation compared with traditional monitoring methods,and has been widely used in the field of civil engineering.However,the existing structural displacement monitoring methods based on computer vision are not fully applicable to the displacement monitoring of seismic isolation bearings during earthquakes.In this paper,we propose a dynamic displacement monitoring method for seismic isolation bearings based on computer vision technology,using a consumer-grade camera with two feature point matching algorithm principles based on SURF feature point optical flow and improved sub-pixel corner point target tracking to realize the dynamic displacement monitoring of seismic isolation The feasibility of the algorithm is verified through experiments.The main research contents and conclusions of this paper are as follows:(1)For the visual monitoring characteristics of the seismic isolation bearing,i.e.,the monitoring equipment is installed in the lower connection layer of the seismic isolation bearing,and considering that the earthquake will cause the camera to identify the measurement error,a special calculation method for the displacement of the seismic isolation bearing is proposed.By placing the camera in the vertical direction normal to the plane of the seismic isolation bearing structure and installing artificial markers on the surface of the connection between the upper and lower structure of the seismic isolation bearing,the visual recognition algorithm can accurately identify the target feature points within the artificial markers installed on the surface of the structure,and then track the pixel displacement information of the feature points within the output markers,and use the conversion of scale factor coefficients to obtain the true displacement data of the upper and lower structure of the seismic isolation bearing respectively,and finally calculate the displacement data of both.The real displacement data of the upper and lower structures of the seismic isolation bearing are obtained by using the conversion of scale factor coefficients,and the relative displacements of the two are finally calculated,which can eliminate the influence of errors and obtain the actual displacements of the seismic isolation bearing.(2)A characteristic optical flow algorithm based on SURF feature point matching is proposed.By identifying the SURF feature points on the artificial markers installed on the structure above and below the seismic isolation bearing,the dynamic displacement monitoring of the seismic isolation bearing is realized with the feature points as the tracking object.The method was verified by vibration test of the vibration isolation bearing,and the following conclusions were obtained: the displacement time range curve measured by the method and the displacement time range curve measured by the laser displacement sensor are in good agreement;the peak error of horizontal displacement of both measurement methods under each working condition is less than 3%;the absolute error of the peak of horizontal displacement measurement is not more than 0.3mm,which meets the accuracy of horizontal displacement measurement of the vibration isolation bearing.This method proves that it is feasible to monitor the displacement of the seismic isolation structure.(3)To address the shortcomings of the SURF characteristic optical flow algorithm,a target tracking algorithm based on improved sub-pixel corner points is further proposed,which can improve the accuracy of characteristic point identification and the stability of target tracking by using characteristic corner points as the target identification tracking object to realize the monitoring of seismic isolation support displacement.The method is verified by monitoring the shaking table test of the vibration isolation bearing of the shaking table model structure and the basic mechanical performance test of the vibration isolation bearing.The test results show that the horizontal displacement results measured by the method in the two tests are in good agreement with those measured by the displacement meter;the peak horizontal displacement errors are less than 1%;the absolute peak displacement errors are no more than 1 mm;compared with the SURF-based characteristic optical flow algorithm,the peak displacement errors of the seismic isolation bearings measured by the method are reduced,and the accuracy and stability are higher.The feasibility and accuracy of the method are proved.