Displacement Monitoring And Modal Identification Of Bridge Based On Computer Vision

Displacement monitoring and modal identification are the key contents of bridge structural health monitoring,which play an important role in bridge structural safety performance evaluation.In order to solve the problems of inconvenient sensor installation and high cost in the traditional contact measurement method,the noncontact measurement method of structure based on computer vision is studied in this paper.Using the Zhang calibration method and the feature point matching method as the basic algorithms,the non-contact displacement measurement system of structure is established combined with the consumer camera,and then it is applied to the identification of structural modal parameters.The main research contents of this paper are as follows :(1)Reading the relevant literature,the principle of computer vision displacement measurement is discussed.On the basis of studying the principle of coordinate system conversion,camera calibration and common target tracking algorithms,the Zhang camera calibration method and the Speeded Up Robust Features(SURF)algorithm are determined as the basic algorithms in this paper.(2)Aiming at the problems of low accuracy,slow calculation speed and instability of existing SURF feature matching algorithm,an improved target tracking algorithm is proposed in this paper.Firstly,sub-pixel SURF algorithm is used to improve the tracking accuracy.Secondly,the tracking stability is improved by matching multiple points,Finally,the kernel correlation filtering(KCF)algorithm is used to realize the two-step tracking process.A search window is defined in advance and it is roughly located by using KCF.Then,image segmentation and sub-pixel SURF are used to further locate the target accurately.Comparing and analyzing the processing results of SURF,KCF and the improved algorithm for the same video,the results show that the improved algorithm proposed in this paper has high speed and accuracy.(3)Based on the improved target tracking algorithm,the image preprocessing,3D reconstruction and displacement calculation algorithm program are compiled and integrated using Open CV Python.The smart phone is used as the sampling device to build the structural displacement monitoring system.Combined with the covariancedriven stochastic subspace(Cov-SSI)method,the identification of structural modal parameters is realized.The feasibility of the algorithm program is preliminarily verified by camera calibration experiment and the processing the video of Humen Bridge.(4)The vibration experiments of cantilever beam and fixed end beam are carried out under laboratory conditions to verify the accuracy of displacement measuement and modal parameter identification of the visual displacement monitoring system.The experimental results show that the system has high accuracy in displacement measurement and can well reflect the dynamic changes of structural displacement.Taking the displacement time history as the response signal,the modal parameters of the bridge structure can be better identified combined with the modal identification method.This paper realizes the application of computer vision principle in structural displacement monitoring and modal identification,and it is verified to be practical.The research can provide reference for subsequent similar research work.