Fusing Optical Flow and Inertial Data for UAV Motion Estimation in GPS-denied Environment
Hou Zhi, Juntong Qi, Mingming Wang · 2019
In this paper, a solution for estimating the motion of unmanned aerial vehicle (UAV) in lacking of GPS environment is proposed. First, interframe difference method is used to detect the point of interesting in adjacent frames. Then Pyramid Lucas-Kanade (Pyr-LK) algorithm is used to calculate optical flow of the point of interesting. Furthermore, mean-shift algorithm is applied to improve the accuracy of optical flow. Finally, the optical flow and Inertial Measurement Unit (IMU) data are fused based on Extended Kalman Filter (EKF) to estimate velocity. Flight test is conducted, and the comparison experiment results of our method with the differential GPS which can provide centimeter-level positioning data show that the proposed method can estimate the motion velocity of UAV precisely.