Motion State Estimation Of UAV Based On Binocular Vision And Inertial Navigation

With the popularity of small and multi rotor unmanned aerial vehicles in recent years,multi rotor UAVs have become a common operation and research platform.Its working environment is limited by artificial restrictions and has a wide range of applications.At present,the multi rotor unmanned aerial vehicle(UAV)is mainly used in the fields of electric power cruise,agricultural spraying,environmental monitoring,and transportation of goods.With the development of robot technology,the requirements of robot vision technology are also increasing.Stereo vision technology is the direction and hot spot in the field of digital image processing and computer vision in recent years.Binocular vision system can obtain abundant environmental information and has a passive measurement,small volume,low power consumption,low cost and other advantages,the development of image processing technology and computer vision theory and the improvement of processor speed of computer vision has been applied more and more in micro motion UAV state estimation,The main idea of estimating the motion state of UAV Based on visual information is to extract optical flow information from image sequences captured by cameras,and then fuse IMU and altimeter to estimate the motion state of UAVs.The main contents are as follows:(1)the UAV platform was designed.After that,the motion state estimation algorithm was transplanted to the four rotor UAV platform.First,the image was captured in real time by camera,then the image processing task was completed in real time using the embedded card computer ODROID XU4 GPU.(2)in the first offline calibration of two cameras that camera parameter matrix of the camera,then the image preprocessing,including grayscale conversion,correction and morphological filtering;depth information by binocular vision system to obtain the optical flow information and the environment,the speed and angular velocity and inertial navigation components of acquisition integration the real-time optical flow information decomposition,finally obtains the UAV motion accurately.(3)the UAV flight platform was built by using the Pixhawk flight control board.The algorithm was transplanted to the ODROIDXU4 card computer,and the algorithm was verified by Opti Track vision location system.