Reconstruction Of Vehicle Environment Based On Vehicle Laser Scanning System

With the rapid development of intelligent driving technology,higher requirements are put forward for the environmental perception ability of vehicles.The ability to reconstruct the environment in front of the vehicle is regarded as the key to improve the perception ability of the vehicle.The vehicle laser scanning system has the advantages of suitable data quantity,small environmental impact,high measurement accuracy and long detection distance.Therefore,the vehicle laser scanning system is widely used in intelligent vehicles to reconstruct the environment in front of the vehicle,In this paper,the vehicle laser scanning system is used to reconstruct the environment in front of the vehicle.In this paper,lidar,GPS and inertial navigator are used to reconstruct the environment in front of the vehicle in real time.The main research work of this paper is as follows:Firstly,the design idea of the vehicle laser scanning system is described,and according to the design idea,an experimental platform for the reconstruction of the vehicle front environment is built.The experimental platform consists of lidar,GPS and inertial navigator.The measurement index,testing principle and error model of each sensor are introduced,and the error of laser scanning system is analyzed.According to the needs of the project and other aspects,a reasonable sensor instrument is finally selected to meet the requirements of testing accuracy.Secondly,time registration of multi-sensor data of vehicle laser scanning system is carried out.In this paper,multi-threaded technology is used to collect multi-sensor data,and Boost network timer is used to mark the time scale of each group of sensor data,so as to ensure that the multi-sensor data adopts a unified time benchmark.At the same time,the interpolation and extrapolation method is used for time registration to reduce the impact of measurement accuracy caused by time error of each group of data.Furthermore,the integrated navigation Kalman filter algorithm is used to calculate the attitude position.There are drift errors in the output data of the inertial navigation system,and with time accumulation,this paper uses the Kalman filter algorithm of indirect output correction to calculate the position and attitude of the output data of the inertial navigation system and GPS,which improves the measurement accuracy of the position and attitude of the integrated navigation system.Finally,point cloud preprocessing and multi-sensor data space registration are completed.Point cloud is sampled down to reduce the density of point cloud,and outliers are removed to improve the accuracy of measured point cloud data.ICP algorithm is used for rough configuration of point cloud,and geometric reconstruction method based on feature surface is used for surface reconstruction of point cloud data.The data measured by lidar,inertial navigation system and GPS are registered in space,and the point cloud data is transformed from lidar coordinate system to geographic coordinate system and then to WGS-84 coordinate system.The multi frame point cloud data is unified into WGS-84 coordinate system through coordinate transformation,and the 3D point cloud map of vehicle front environment is constructed to complete the real-time reconstruction of vehicle environment.