Research On Bridge Structure Deformation Measurement Based On Digital Image Correlation Technology

As the economic lifeline of China,bridge construction plays an inestimable role in the development of our country.Bridge structure deformation detection plays an important role in the process of bridge construction and operation management.At present,the bridge structure deformation monitoring is mainly based on traditional manual measurement methods,such as total station,level,contact displacement meter,etc.,but the manual measurement methods have limitations,and the accuracy is greatly affected by human,time-consuming and laborious.With the development of visual image technology,high accuracy,low cost,non-contact computer vision measurement methods are widely used in many fields.In this paper,the application of computer vision testing methods in the field of bridge inspection is studied by using mobile mold frame field construction testing and laboratory testing of reinforced concrete simply supported bridge models.The details of the study are as follows.(1)Through the understanding and in-depth study of the theoretical knowledge of digital image related technology,the generation of digital image related technology and its principles are elaborated,and the matching algorithm between before and after digital image deformation is briefly explained.The ideal imaging model of the camera,small-aperture imaging,is introduced,and the four coordinate systems are used to visually describe the conversion between two-dimensional and three-dimensional coordinates,and the relevant equations are used to derive the conversion process from two-dimensional to three-dimensional.The aberrations in the actual camera model are illustrated.(2)Develop and design a monocular image bridge deflection monitoring software applicable to this paper and conduct field tests.Based on the basic theoretical principles related to monocular vision,the whole monitoring software is designed by integrating the improved ROI automatic extraction key algorithm.The software can automatically extract the main girder rivet plate area from the main girder images collected in the field,and then perform distance calculation based on the scale factor obtained from the active calibration;the software is used in combination with commercial monocular video monitoring software to monitor the deflection changes during the pre-pressure of the bridge moving formwork in the field.(3)Overall design and laboratory testing of binocular vision measurement software.Based on Python language combined with binocular vision measurement principle to develop the whole software,the software design mainly includes image acquisition and pre-processing,stereo calibration,stereo matching,depth calculation and 3D reconstruction.The software design includes image acquisition and pre-processing,image pre-processing,stereo matching,depth calculation and 3D reconstruction.The sub-pixel gray-scale interpolation is introduced to ensure the test accuracy,and the stereo calibration is based on Zhang Zhengyou’s checkerboard grid calibration.The results of the field and laboratory experiments show that the deflection monitoring data of the monocular image measurement system during the pre-pressure of the bridge moving formwork is comparable to that of the commercial monocular video measurement software,and the error range is kept within a controllable range,and its measurement results are more accurate compared with the traditional level,and the measurement method is more time-saving than the manual measurement.The vertical displacement measurement results of the binocular measurement system in the concrete simply supported bridge bending experiments are similar to the traditional measurement results.The binocular vision system can collect data for the whole process of concrete beam damage,and the commercial software Match ID can be used for the full-field displacement and strain analysis calculation in the scattered area of concrete simply supported bridge.The feasibility and superiority of digital image-based correlation techniques are verified in both outdoor field and indoor laboratory environments with large fields of view.