Research On Oblique Photogrammetry Technology Of Bridge

As a big country in bridge construction,monitoring,testing and evaluating the structure status of bridge structure is a problem that needs to be solved in daily maintenance.Therefore,the efficiency of traditional bridge detection is low.Therefore,the intelligent detection technology of bridge comes into being.However,the incomplete design and construction data of some old Bridges in practical projects hinder the development of intelligent detection technology of Bridges.To meet the needs of acquiring structural information of these Bridges,this thesis uses oblique photogrammetry technology to acquire digital information model of Bridges.With the advantages of fast,real and spatial geographic information,it can provide technical and data support for bridge information management and BIM reconstruction.In this thesis,the specific bridge engineering structure is taken as the experimental research object,and the oblique photogrammetry,computer vision and other technologies,as well as the hardware platforms such as UAV and computer are used to study the key issues of the 3D reconstruction of the bridge in real scene and the following work is carried out:(1)The key technologies and basic principles of oblique photogrammetry and 3D reconstruction are studied,including the principle of photogrammetry,image joint adjustment,scale invariant Feature conversion(SIFT)algorithm involved in image intensive matching,multi-view stereo reconstruction & Structure From Motion(SFM-MVS)algorithm,point cloud network,texture mapping and other technologies.(2)In view of the special structural form of bridge structure,a specific UAV platform and flight control software are selected to form an acquisition system for field work in the early stage,and the calibration principle of non-measurement camera is studied in depth.Therefore,the calibration work of UAV camera is carried out,and the relationship between camera pixel position and the real field scenic spots is established.The corresponding photogrammetry and 3D reconstruction software and the computer platform are selected to form the internal processing system,and the software and hardware system of the bridge oblique photogrammetry is established.(3)Based on the technology and principle of photogrammetry and 3D reconstruction,the experimental code of feature extraction and matching and the construction of Delaunay triangulation network were written in Python language,combined with the source code of MVS algorithm and Context Capture software,the experimental verification of 3D reconstruction of the captured stone lion and chair image sets was carried out.(4)Study the image acquisition and reconstruction method of bridge 3D model based on UAV oblique photogrammetry and proximity photogrammetry technology.Combined with this method,flight path design and image collection are carried out for the actual bridge projects such as road and bridge in a city of Wanzhou,Tuojiang Simei Bridge in Neijiang and Qijiang Chengbei Bridge.The collected images are checked and pre-processed.This thesis explores the technical process of establishing the real 3D model of bridge structure through Metashape and Context Capture software,and then successfully establishes the real 3D model of bridge.After trimming the real 3D model,the appearance refinement of the results is evaluated according to the specifications.The point cloud classification of bridge structure is studied to realize the monomer of bridge structure.The cloth filtering algorithm is used to separate the bridge body from the ground point and carry out the comparison and data analysis with other filtering methods.(5)The model accuracy and structural accuracy were evaluated according to the results of the real 3D model based on UAV oblique photogrammetry.The results showed that the 3D model of the bridge reconstructed with this method met the accuracy requirements in the specifications.Finally,based on RANSAC algorithm,the structure and form composition of point cloud model of individual bridge was detected and automatic size extraction of bridge structure was realized.The data of automatic extraction of bridge structure information was analyzed.