A new scheme of vision based navigation for flying vehicles-concept study and experiment evaluation

In this paper, a new scheme of vision based navigation was proposed for flying vehicles. In this navigation scheme, the main navigation tool is a camera, plus an altimeter. The feasibility of this navigation scheme was carefully studied both from theory and numerical analysis. Unlike most of vision based navigation approaches using feature trajectories to compute 3D-platform motion, we use the image geometrical transformation parameters between consecutive frames to infer3D displacement of camera. Due to this change, the navigation process can be conducted even if there is no salient features that can be extracted from in the image sequence, for example, in the case of flying over the sea. As a result, the long-range navigation becomes possible by use EO sensor. Moreover, the way of improvement navigation accuracy was also discussed in the paper. The experiment results demonstrated that the navigation accuracy of this system is compatible to GPS (Global Positioning system), much higher than all kinds of INS (Inertial Navigation System) in terms of position estimation. It is a good alternative choice when the GPS signal is not available.