Research on Point Cloud Registration Between Low-altitude Unmanned Aerial Vehicles and Ground-based Mobile LiDAR
The matching and fusion of multi-platform LiDAR data is currently a hot research topic in the field of point cloud applications.This paper takes the registration between unmanned aerial vehicles(UAVs)and ground-based mobile LiDAR as an example.By collecting point cloud data of four types of terrain:built-up areas,sparse forests,dense forests,and regular forests,this study employs manual registration,ICP registration,and a combination of both methods to investigate the applicability of different registration methods to various data scenes and analyze the factors affecting point cloud registration accuracy.The experiments show that in terms of registration accuracy,built-up areas＞sparse forests＞dense forests＞regular forests,indicating that the stronger the geometric characteristics of the study area,the better the registration effect.In terms of registration methods,the method of combining manual registration with ICP outperforms single registration methods,achieving better point cloud matching in areas with less prominent features,which is more applicable for data registration of forest regions.
unmanned aerial vehicle LiDARground-based mobile LiDARpoint cloud registrationICP algorithm
NETL
NSTL
万方数据
