Research On Structural Dynamic Response Monitoring Method Based On Image Sequence

Dynamic parameters are one of the important parameters in the health monitoring of civil engineering structures,and they are the dynamic response of the structure under external loads.By monitoring the parameters,you can understand the working status of the structure in time,and provide a basis for structure performance evaluation,damage detection,reliability status evaluation and maintenance,etc.At present,a variety of methods for obtaining structural dynamic parameters have been developed,and there are many disadvantages in the traditional contact measurement method,such as: difficult installation,short life,complicated wiring,and high cost;the emerging non-contact methods such as dynamic laser rangefinders,total stations And target-based image processing technology have disadvantages such as high cost,susceptibility to electromagnetic interference,and difficulty in target placement.In order to solve the above problems,it is necessary to develop a low-cost and more efficient structure dynamic response monitoring method.Based on this,this thesis combines digital image processing technology and civil engineering structure monitoring technology to propose a structure dynamic response monitoring method based on image sequence.The main work is as follows:A dynamic displacement monitoring program based on image sequence is compiled by Python,which achieves the selection and matching of any structural measurement points of interest within the image.The optical flow method is used to calculate the pixel displacement of feature points between consecutive images to make the calculation at a sub-pixel level,which improves the accuracy of displacement monitoring.Combining the principle of fast Fourier transform,using Python to write a program to achieve the extraction of structural vibration frequency.Aiming at the problem of inaccurate tracking of structural measuring points in some videos,the feature point drift is limited by adding ROI area,and the template matching is combined with the Shi-Tomasi feature point algorithm to write a feature point tracking program based on region matching,so that the program can stabilize track the feature point in each frame of the sequence.The motion curve and the pixel coordinates of the feature point can be exported in each frame.A dynamic displacement monitoring system based on image sequences was built,and a small cantilever beam in the laboratory was used to study the influence of different video collection distances and different video collection frame rates on the monitoring results,and the cable-type displacement sensor was obtained at a video collection distance of 17 m.The very consistent results indicate the accuracy and feasibility of the method.