| Recently,with the increase in the number of vehicles and the mileage of road traffic,the speed of road pavement damage has accelerated,and many different types of diseases have gradually appeared on the pavement.Pothole is a typical disease of asphalt pavement,which seriously affects the pavement smoothness,driving comfort and road performance.Traditional investigation methods based on artificial vision are time-consuming,labor-intensive,low-precision,and high-risk,which is already difficult to meet the needs of modernization and scientific maintenance.As a new spatial data collection technology,the vehicle-borne mobile measurement system can quickly collect large-area three-dimensional spatial data,which provides data support for the extraction of Pothole.At present,the difficulties in extracting pavement pothole based on vehicle-borne laser point clouds are:the point cloud has a large amount of data and redundant data exists;the discrete distribution of point clouds is difficult to model;the shape of the pits and grooves is variable and cannot be expressed with a single parameter.On the basis of analysis of the pothole shape characteristics,this paper studies an extraction method of pothole information in vehicle-borne laser point clouds.The main contents are as follows:(1)It combs the development status and trend of pavement breakage detection system and vehicle-borne mobile measurement system,expounds the data collection process and the formation mechanism of vehicle-borne mobile measurement system,and analyzes in detail the advantages and feasibility of laser point clouds extraction of the pothole.(2)According to the spatial distribution characteristics of road pavement and boundary point clouds,a pavement point clouds extraction method that takes into account the pavement shape is studied.It constrains the plane position and elevation of the driving trajectory to establish a spatial buffer to effectively reduce redundant data;constructs a plane grid symmetrical to the driving trajectory so that the same features exist in the same grid;performs point clouds within the grid elevation layering to remove noises such as road vehicles and debris;uses the difference in elevation between the grid and adjacent grids to filter the pavement grid,and combines with the regional growth algorithm to obtain a complete and accurate pavement point cloud for extraction of pothole basic data is provided.(3)In-depth analysis of the plane shape and section shape of pavement pothole,an automatic extraction method of pothole information based on the description of the curvature characteristics of the sectional line is studied.The curve fitting equation of the profile line is established to realize the model expression of the point cloud.The characteristic points of the high curvature are used to describe the edge of the outline of the pothole.The angle between the characteristic points on the profile line is used as a constraint to extract candidate boundary points.Clustering and shape-constraining denoising of candidate boundary points,filtering out isolated points,cracks and subsidence diseases,and obtaining accurate information of pothole contour,depth,area,location and profile morphology. |
