Research On Road 3D Modeling Method Based On Vehicle LiDAR Data

With the rapid development of economy and society,the infrastructure of national production and life has been more improved,and the road,as the link of the city,have reached unprecedented levels of coverage.Road detection for built roads is the first priority for applications such as rapid road maintenance,digital city construction,and assisting emerging intelligent driving.Emerging vehicle-based mobile measurement technologies have unique advantages in quickly and accurately acquiring object-target information.Applying on-board mobile measurement technology to road measurement,deformation monitoring,etc.,can quickly realize 3D visualization of roads and improve road maintenance efficiency.Has a good research significance and application prospects.This paper studies three-dimensional modeling of roads based on vehicle-mounted LiDAR data,processes the point cloud data collected by the vehicle-mounted mobile measurement system,and finally realizes the construction of a road model.The sensors used in the vehicle-mounted mobile measurement system used in this article mainly include three-dimensional laser scanners,GPS navigation systems,INS inertial navigation systems,and CCD cameras.In this paper,the working principle of the vehicle-mounted mobile measurement system and the process of field data collection are introduced in detail.The processing of point cloud data is mainly studied,and the construction of a road three-dimensional model is completed through the processing of measured data.For the point cloud data processing,this paper first introduces the current advanced professional point cloud data processing method library PCL,and converts the collected road point cloud data into coordinates and other formats.Through human-computer interaction,pass-through filtering,statistical filtering and radius filtering,the filtering of point cloud data is gradually realized.For road voids formed after filtering,a sample data set is taken above and below the hole to repair the hole,which improves the accuracy of repair.The road boundary is an important feature of the road.It focuses on the extraction of the latitude and longitude line boundary.Combined with the PCL’s normal vector judgment,the Open MP acceleration method can be used to achieve normal line estimation in the PCL point cloud library,remove non-boundary points,and improve the boundary extraction.Accuracy.The downsampling of the road point cloud data through the PCL’s VoxelGrid,under the premise of retaining the characteristics of the point cloud,removes the redundant data,achieving a very good streamlining effect.Finally,the three-dimensional model of the road is constructed by using the Greedy Projection triangulation method of the PCL library for the processed measured data,and three-dimensional visualization is performed for each stage.The engineering application shows that the accuracy of LiDAR data acquired by vehicle-mounted mobile surveying can meet the requirements of road survey and survey,and the generated road three-dimensional model can be applied to smart cities,unmanned vehicles,road maintenance road landscapes and other projects.