ビジョンセンサを用いたマルチコプタの自律飛行

抄録

In this paper, we describe the control of an unmanned aerial vehicle (UAV) using a vision sensor with visual markers. Our research has been developing a UAV that performs autonomous flight to inspect concrete bridges. In general, UAV relies on GPS for self-position estimation. However, UAV can not acquire GPS signal in our objective environment, under bridges. Hence another sensor is necessary to estimate the position in an environment where GPS can not be acquired. Therefore our UAV use a vision sensor to estimate its own position and control it. In this study, we employs ArUco algorithm to estimate UAV’s position referring visual markers. There is still a possibility that the UAV miss the markers and can not control its position. Therefore we implemented an algorithm referring other sensor during the missing situation. An experiment shows that the UAV can control its position using our algorithm.