UAV Simultaneous Localization And Mapping Based On Vision

UAV has become a hot research topic in recent years.It has high use and research value both in military and civil.Accurate estimation of UAV's position and attitude is the key to achieve autonomous flight of UAV.At present,the main method is to adopt GPS(Global Positioning System)and inertial navigation system to achieve the UAV navigation.But in the actual indoor environment,GPS signal can not be got and we need to use computer vision to locate the UAV indoor.Therefore,this paper studies the control method,simultaneous location and mapping,which is based on the need of the autonomous flight of the indoor unmanned aerial vehicle.The main work is as follows.First of all,the development status and key technologies of UAV are reviewed,including the UAV flight control system and the application of computer vision technology in UAV.Secondly,establish the dynamic model of the UAV and the UAV flight control system design.Complete the simulation in MATLAB and the development of flight control procedures based on ARM controller.The semi physical simulation system is established to verify the control method.Then,the algorithm of visual simultaneous location and mapping based on RGBD camera(RGBD SLAM)was studied.Under the Linux system,the RGB and depth images acquisition and processing of the computer platform is realized.The estimation of the position and pose of the UAV body,and the construction of the 3D scene of the system are realized.At last,the RGBD SLAM algorithm is improved,and the new RGBD SLAM algorithm is obtained.And complete the experiment of UAV airborne simultaneous location and mapping.