Research On Measurement Method Of Bridge Dynamic Displacement Based On Computer Vision

In the filed of bridge health monitoring,displacement measuring is an important method for structural damage identification and condition evaluation,and an important means for structural mechanical analysis.With the development of bridge construction technology in China,the number of Bridges has increased,and various types of bridges have been in serve.However,due to the continuous load sustained during the service period of the bridge structure and the occurrence of deterioration and damage,the research of bridge engineering has gradually shifted from construction to maintenance.In the bridge health monitoring system,it is necessary to select appropriate sensors to collect various physical indexes according to the site conditions.In the displacement test,the traditional sensors show many disadvantages,such as large number,difficult installation,high price,wear and tear,etc.They limit their use in practical engineering.New technologies are needed to solve these problems.In recent years,with the rapid development of artificial intelligence technology represented by deep learning and computer vision technology,great breakthroughs have been made in the field of target tracking and image recognition.The continuous penetration and development of computer technology in the field of engineering also provides a new idea for the development of displacement measurement technology.This paper uses the computer vision technology to develop the displacement testing technology,and presents a new effective displacement testing method.Aiming at the problems existing in related researches,the algorithm is further optimized.The influence of the resolution in the actual application of the algorithm is analyzed.And developed the software user interface.The main research work of this thesis is as follows:(1)In the displacement testing method based on optical flow algorithm,non-local mean denoising technology and ROI relocation process are adopted to achieve higher precision testing.While keeping the clear contour of the image,the noise information can be effectively suppressed,and the balance of noise reduction scale and edge clarity is well balanced.The test method proposed in this paper shows higher accuracy in the displacement test of model bridge,effectively reduce the phenomenon of displacement drift,and is less sensitive to the subjective parameters.In frequency domain information,the testing techniques based on computer vision show performance comparable to that of traditional testing methods.(2)In the vehicle load test,this paper verified that the displacement test method based on optical flow algorithm has good robustness under the condition of resolution change and can achieve stable accuracy under different video resolutions by analyzing the test accuracy under multiple resolutions.An algorithm that is more robust to resolution can be used to perform faster calculations in smaller video volumes and achieve an acceptable level of accuracy(3)This paper develops the user interface for the tedious configuration of the code running environment.After comparing several user interface development schemes,Piside2 is used to develop the Python code implemented in this study,and good results are achieved.After using the user interface to conduct displacement testing,both the testing efficiency and the visualization level are improved.