Vision-inertia based navigation algorithms for quadrotors

Abstract

In this paper, the navigation problem is investigated for quadrotors which fly in a GPS-denied indoor environment with ground grids. A vision-inertia based navigation algorithm is proposed which only uses the information of the grid intersection points and inertial measurement unit. First, the grid lines are detected by merging pixels with adjacent location and approximate intensity. Characteristic parameters of the lines are obtained by solving a least square problem. Then, the locations of the intersection points are calculated using these characteristic parameters. Secondly, the intersection points are matched with the corresponding points located in the real environment by solving a joint optimization problem, meanwhile the optimal homography matrix is obtained. Thirdly, the navigation information is estimated by an Extended Kalman Filter which uses the pose recovered from the homography matrix as measurements. Finally, the experiment results show the effectiveness of the proposed navigation method.