Research On Potential Damage Locating Technique Of Bridge By Gaussian Curvature With Single Phase Point Cloud

In recent years,with the development of infrastructure construction and traffic network in China,the use of Bridges is increasing.However,due to the comprehensive influence of service life,environment,overload,traffic congestion and geological activities,the safety situation of bridge operation becomes increasingly serious,security risks have increased significantly.It is urgent to carry out bridge damage identification and potential damage area detection.The emergence of ground three-dimensional laser scanning(TLS)can transform the limited control "point" into a solid "area",comprehensively obtain a large number of coordinate point information on the bridge surface,quantitatively analyze the overall damage of the bridge and locate the potential damage area through the comparison of multi-stage surface models.However,three-dimensional laser scanning measurement has randomness and non repeatability of multiple scans.In view of the above problems,a set of potential damage area detection method of single-phase point cloud integrating Gaussian curvature of bridge deck and inclination of pier is established to evaluate the potential damage of the whole bridge.This research can provide basis for accurate flaw detection and promote the rapid and accurate development of bridge safety monitoring and maintenance.The details are as follows:(1)It is difficult to accurately obtain the potential damage area of the bridge in a single phase point cloud,according to the uniform and continuous bending theory of the curved surface before and after deformation under load,the feasibility of the potential damage detection of the bridge deck by the Gaussian curvature of the point cloud is analyzed in detail.A normal vector constrained Gaussian curvature coordinate transformation damage detection method is adopted.By constructing the lower surface model of the bridge,the single phase point cloud data is transformed into the coordinates of normal vector constrained Gaussian curvature,the Gaussian curvature(bending degree)of the lower surface point cloud of the bridge is calculated,the potential damage area is delimited through the mutation of the value of Gauss curvature,and the damage area is divided into grades.The experimental results show that the single-phase point cloud of the coordinate transformation method of normal vector constrained Gaussian curvature has a certain noise suppression effect,can retain the characteristic information of the bridge model,the extreme point of Gaussian curvature can locate the location of bridge deck damage,improve the judgment standard of bridge deck damage area,and provide accurate services for bridge damage judgment.(2)In view of the different stress in each area of the bridge deck,the conventional pier perpendicularity detection only calculates the horizontal offset for the two sections of the pier bottom and the pier top,and can not detect the spatial alignment of the central axis of the pier.A tilt detection method combining slice method and horizontal section method is proposed to extract and segment the horizontal boundary of pier column.Based on the idea of straight line fitting in mathematical geometry,fit the center of point cloud slice and pier column axis,calculate the vertical deviation of pier column axis,and analyze the tilt trend of pier column and its influence on bridge deck through the variation law of vertical deviation.The experimental results show that the combination of the two methods can quickly and accurately obtain the variation of pier inclination with height,and meet the accuracy requirements of pier verticality observation.At the same time,it can effectively distinguish and eliminate the inclination or inclination illusion caused by the initial inclination of bridge pier column construction,the settlement of bearing platform or foundation pile,the action of external load and surface concave convex phenomenon.