Research On Non-target Measurement Method Of Bridge Displacement Based On Image Matching Technology

Displacement is an important index of bridge performance evaluation and health evaluation,which directly reflects the overall stiffness of the structure.Traditional bridge displacement measurement methods can be divided into contact and non-contact.Contact displacement measurement method needs to connect wired or wireless sensors to the structure or component.A fixed reference point is required.This method has high cost and can only measure one location.The non-contact measurement methods,such as total station,laser vibration meter,GPS measurement,etc.,have the disadvantages of expensive measurement equipment,low measurement accuracy and difficult to realize high-frequency continuous automatic measurement,so they are not suitable for bridge micro dynamic displacement measurement.In this paper,a bridge displacement measurement method without target based on image matching algorithm is proposed,and a displacement measurement system including software and hardware is constructed.This method does not need to install special marks on the bridge,and can measure the displacement of multiple positions of the bridge structure synchronously,quickly and accurately.The reliability and accuracy of the displacement measurement system are verified by laboratory test and real bridge test.The main contents and conclusions are as follows:(1)The bridge feature extraction,description and matching algorithm based on image matching technology are studied.The advantages and disadvantages of existing image matching algorithms in feature detection,description and matching are analyzed.According to the requirements of bridge displacement measurement,a fast,stable and accurate bridge image feature detection and matching method is proposed by combining surf algorithm and freak algorithm.The PROSAC algorithm is used to optimize and purify the matching points and eliminate the wrong matching.In view of the fact that the collected image contains more useless information,an automatic extraction method of ROI in grid format is proposed to reduce manual operation and improve measurement efficiency.(2)The bridge displacement transformation method based on image matching technology is studied.The basic principle of camera calibration is studied,and Zhang Zhengyou chessboard calibration method is used to calibrate the camera internal parameters and distortion parameters.The relationship between pixel coordinates and physical coordinates is established.For practical application,the scale factor method is used to simplify the extraction process of physical displacement.(3)The bridge displacement measurement system based on image feature matching algorithm is constructed.The accuracy of displacement measurement system based on image feature matching technology depends on many parameters.A method to evaluate the accuracy of displacement measurement system based on displacement measurement error distribution of static structure is proposed.When the camera is placed at different distances to measure the displacement of the static structure of the laboratory,the measured non-zero displacement must be an error.SPSS software is used to verify that the error distribution is normal distribution with mean value of 0.The interval of 2 times standard deviation of error distribution is defined as system precision.The relationship between precision and distance of displacement measurement system is fitted.(4)The robustness of the displacement measurement system under different illumination intensity and camera vibration is discussed.In the range of 2.5 lux to 1000 lux,the displacement of the top floor of the frame under the impact excitation was measured.The results show that when the illumination intensity is less than 30 lux,the accuracy of displacement measurement decreases significantly.By simultaneously measuring the displacement of the static reference point in the background,the effectiveness of eliminating the influence of camera vibration by subtracting the displacement of the static reference point is verified.In addition,considering that there is no suitable static reference point in the field of view of the camera in the actual measurement,a method is proposed to eliminate the camera vibration by using two cameras installed on the same tripod to measure the static reference point and structural displacement respectively.(5)The feasibility of multi-point displacement synchronous measurement and modal identification is verified.Under the horizontal excitation of knocking on the top floor of the frame,the displacement of each floor is measured simultaneously by using the displacement measurement system.Based on the displacement data,the modal characteristics of the four story steel frame,including natural frequencies and modes,are identified by using the stochastic subspace method.The results show that the first three natural frequencies and vibration modes identified by the displacement data measured by the displacement measurement system are consistent with those identified by the laser displacement meter and accelerometer.(6)The effectiveness and feasibility of the displacement measurement system based on image feature matching technology are verified by laboratory test and real bridge test.The results of laboratory shaking table test,four story frame vibration test and real bridge measurement of expressway bridge,subway viaduct and pedestrian overpass show that the displacement measurement system can achieve fast and accurate measurement of bridge displacement.