Research On Bridge Structural Damage Identification Method Based On Computer Vision

The performance of bridge deteriorates gradually due to the influence of load and material deterioration during service.It is very necessary to ensure the structural safety of the in-service bridge by means of inspection,evaluation and monitoring.The static displacement(or deflection)and dynamic displacement of the whole bridge are one of the indexes that can best reflect the safety of the bridge structure.Machine vision has the potential of long-term monitoring or short-term measurement of multi-point displacement on bridge surface due to its advantages of non-contact,long distance and large field of view.Aiming at the existing problems of bridge displacement measurement methods based on machine vision,this paper proposes a self-damping dynamic displacement measurement method based on SURF algorithm.SURF algorithm was used as the basis of the method,and RANSAC algorithm was integrated to eliminate the error matching of the feature point algorithm.Local sub-pixel algorithm was further used to improve the accuracy of the algorithm and the efficiency of the algorithm was improved through grid partitioning.In order to measure the whole surface displacement of the bridge underside,the optical path correction of oblique camera was carried out,and finally the whole surface displacement of the bridge could be obtained by interpolation method.In addition,the self-proposed front and rear dual-lens device can effectively eliminate the measurement error caused by the camera’s self-vibration,and solve the vibration problem faced by the vision sensor in outdoor applications.The whole system is verified by laboratory test,which proves the accuracy of the whole structure displacement measurement system and the effectiveness of the dual-lens system in eliminating the camera’s natural vibration.The proposed visual sensor system is used to investigate the lateral connection damage,which is often neglected in the current damage identification research.Transverse connections sometimes do not affect the stiffness of the bridge as a whole,which is more difficult to find than beam stiffness damage.The finite element model is used to analyze the influence of the damage position of the transverse connection on the structure,especially the vibration displacement.Based on the proposed visual sensor which can obtain the whole surface displacement of the bridge,the actual measurability of bridge transverse connection damage is investigated by laboratory scale model test.The experiment proves that the traditional acceleration vibration pickup and conventional vibration characteristic parameters are insensitive to the transverse damage connection of the structure,and the displacement sensor which can recognize the whole surface displacement is effective in the transverse damage measurement.