| Before the mature development of the airborne LiDAR technology,aerial photography image processing technology is widely used in production mapping in the field of surveying and mapping.The mapping theory of aerial photogrammetry has been basically improved.Through a series of image processing methods,the azimuth information of aerial images can be calculated,the position and posture of each aerial image can be restored,and the aerial image-based digital map can be obtained by using the classical 3D reconstruction technology.Terrain restoration based on aerial camera measurement can obtain continuous surface texture information,but it has limitations such as complicated calculation,low precision,and vulnerability to environmental influences.In recent years,with the development of airborne LiDAR technology,its innate advantages in directly acquiring surface three-dimensional coordinates have caused a revolution in the field of surveying and mapping.As a new generation of geodetic measurement technology in the field of surveying and mapping,airborne LiDAR has the characteristics of initiative,high efficiency,accuracy,small weather impact,multiple echoes,etc.,and has the advantage that aerial photogrammetry technology can't match.However,the point cloud data generated by the airborne LiDAR technology has the characteristics of discreteness,density,etc.,and due to its impossibility to obtain the real texture information of the surface,it is weak in the classification and recognition of the features,which is the advantage of the image technology.Therefore,it is of great significance to combine the airborne LiDAR technology with aerial photogrammetry technology to complement each other and make up for the limitations of each measurement technology.The research goal of this paper is the fusion method of airborne LiDAR point cloud and aerial image point cloud.Based on airborne LiDAR technology and aerial photography technology,an airborne LiDAR mapping system was designed and implemented,and we studied the fusion algorithm and its specific implementation of LiDAR 3D point cloud and aerial image point cloud for the purpose of 3D point cloud with real texture information.This paper first designs and implements a set of high precision,autonomous airborne LiDAR hardware system based on embedded technology.Using the Raspberry Pi compute module as the control core,the system synchronously acquires laser scanning data,aerial image data,imu posture data,and global navigation and positioning data.At the same time,a ground station cloud processing software system is designed and implemented,which can remotely control and receive data on the airborne LiDAR hardware system,and generate ground point cloud in real time,supporting point cloud filtering and point cloud splicing.Secondly,this paper analyzes the mathematical model of airborne LiDAR point cloud data generation and the imaging model of the camera in detail,and on this basis,the least square method is used to check the installation error of laser and camera to improve the fusion precision.Finally,this paper proposes two fusion methods,one is to map the ground point cloud to the aerial image according to the camera imaging model,to obtain the real texture information for the 3D point cloud,and the other is based on the 3D image point cloud generated by the sequence image.Using the idea of point cloud splicing,the LiDAR point cloud and the image point cloud are coincident and spliced,and then fused.In this paper,the actual flight scanning experiment is used to verify the proposed fusion algorithm by using the software and hardware system designed in this paper,and the fusion precision of each fusion algorithm is calculated.The final result shows that the fusion algorithms based on the airborne LiDAR laser point cloud and image points achieve decimeter-level fusion accuracy. |
PDF Full Text Request