Bridge Deformation Based On Point Cloud Data

The safety monitoring of bridge structure has always been the focus and difficulty in engineering research Deformation monitoring technology is mostly single-point measurement.When monitoring large bridges,there are some problems such as complex operation,difficult measurement and long working cycle The application of three-dimensional laser scanning technology breaks through the limitation of monitoring means,and can measure the whole building with high precision through non-scanning In this paper,three-dimensional laser scanning measurement is used to obtain bridge point cloud data,and point cloud data processing algorithms are used to extract the detailed features of bridge geometric structure and quantify the deformation of bridge structure.The research results provide a new means for bridge safety monitoring and monitoring.The main research contents of this paper are as follows:(1)Aiming at the problem that the 3D laser scanner cannot obtain the complete point cloud data of the target at one time,this paper constructs a multi-view point cloud registration model based on dual quaternion description under the constraint of plane characteristics according to the geometric structure characteristics of the building/structure itself,proves the effectiveness of the model through simulation experiments,and uses the metro point cloud data disclosed by Wuhan University and the bridge point cloud data collected by itself to realize the stitching of bridge point cloud data of different scanning stations through the multi-view point cloud registration model.(2)Because of the complexity of the whole bridge structure,a new bridge point cloud segmentation method based on bridge spatial context information is proposed to extract the interesting point cloud data of bridge components from massive point cloud data.Based on the spatial distribution characteristics of bridge structure from top to bottom,the probability density distribution in the vertical direction is calculated by kernel density estimation for the point cloud data segmented by octree,and the point cloud containing different component boundaries is screened by combining the distance between points and the normal relationship between points.The point cloud data of pier and the point cloud data of beam bottom are segmented from the original point cloud by combining the distance and normal relationship between points.(3)Due to the unrepeatability of 3D laser scanning and the disorder of point cloud data,it is difficult to determine the monitoring standard for deformation monitoring based on point cloud data,and a bridge deformation analysis method based on bridge deformation experience is proposed.This method analyzes the engineering deformation law of different components of the bridge to determine the point cloud deformation monitoring benchmark of different components: for the pier,the geometric center of the section surface of the pier is determined to carry out multi-temporal pier point cloud data registration and pier verticality detection,and realize the deformation monitoring of the pier with the point cloud line primitive as the monitoring reference.For the bridge superstructure under load,according to the principle of stable and similar distribution of the point cloud virtual reference center of mass,the multi-temporal beam bottom point cloud data registration is carried out through both ends of the beam,and the distance between the multi-temporal beam bottom point clouds after registration is used as the deformation of the bridge superstructure,and the deformation monitoring of the bridge superstructure with the point cloud point primitive as the monitoring reference is realized.The experimental results show that the proposed method can reflect the deformation of different components of the bridge,which provides strong support for the deformation analysis of the bridge using point cloud data.