Research And Design Of UAV Localization System Based On The Fusion Of Binocular Vision And IMU

Along with the rapid development of drone technology,UAV applications are gradually diversified,not only the military field,widely used in aerial photography,inspection,UAV logistics and other professional fields.I believe that,in the near future,drones will play an increasingly important role in people's daily lives.For UAVs,autonomous positioning technology has always been one of the core capabilities of UAVs.It is the premise and basis for UAVs to achieve avoidance,navigation and path planning.In this paper,the binocular camera and the inertial measurement unit(IMU)are used for data fusion,which provides accurate position,attitude and speed information for the drone without GPS or GPS signal is weak.And builds a complete set of Autonomous positioning system for drones that is integrated binocular camera with IMU.The main work and achievements of this paper are as follows:(1)Elaborating the fundamental theory of binocular camera calibration,IMU inherent parameter calibration based on Allan variance and Coordinate relationship calibration of camera and IMU.Designeing an evaluation index of binocular camera calibration and Coordinate relationship calibration of camera and IMU.Completing the calibration experiment,geting the relevant calibration parameters and checking the error of calibration results.(2)Introducing the basic flow of binocular vision positioning based on optical flow tracking.The basic principles of various corner detection,optical flow tracking,mismatched point removal,and camera motion estimation are studied.The algorithm design of FAST corner uniformization is completed.With the help of IMU,the optical flow tracking algorithm and the inter-frame outer point removal algorithm are improved.Experiments show that the improved algorithm improves the computational efficiency.(3)Designing a set of binocular IMU fusion positioning algorithm based on the method of Nonlinear optimization.On the Euroc dataset,our method is compared with the current mainstream visual and IMU fusion positioning algorithm.The results show that,in terms of positioning accuracy and computing load,the algorithm presented in this paper performs best.(4)Autonomous positioning system for UAV based on binocular and IMU fusion was built.In the actual scene,experimental verification research was carried out.Experiments show that the positioning error of proposed algorithm is approximately 1.6%,basically meets the requirements of UAV positioning without GPS signals.