Visual navigation for UAV using optical flow estimation

This paper proposes a visual navigation system for an unmanned aerial vehicle using optical flow in a GPS-denied environment. The optical flow of sequence image which is taken by monocular camera is based on block-matching algorithm, An extended Kalman filter fusing the IMU date and pressure sensor measurements is applied to estimate global position and velocity of UAV. This system applies to UAV whose altitude variation is not very large. In addition, the simulation results show that ideally the system can provide high accurate position and velocity, but the accuracy will reduce with the increase of altitude variation in longitudinal plane.