A direct location method based on the combination of spatial information and video

DOI:	10.11809/bqzbgcxb2024.07.006
Keywords:	drones; spatial information;video integration; positioning
Abstract:	Unmanned aerial vehicle remote sensing has the advantages of flexibility, low cost, and strong timeliness, and has become an important means of obtaining ground target image information. The accurate positioning of targets plays a crucial role in target recognition, image processing, and analysis. To achieve the practical requirement of high precision restoration of target location information such as buildings and landmarks in the actual physical environment in a virtual experimental environment, a drone geographic video data model suitable for 3D GIS is constructed. Through hardware level synchronization, the drone’s position, attitude information, and video are fused to establish a corresponding relationship between video data and spatial data. The target is positioned using techniques such as drone flight position, attitude restoration, and video matching. Multiple sets of calculated video positioning data are compared with GPS field collected data. The positioning accuracy with a maximum deviation of 5.772 m and a mean square error of 3.817 m has been achieved, which basically meets the target positioning requirements of unmanned aerial vehicles.
Published:	2024-07-26
PDF