Efficient Aerial Image Matching Algorithm for Autonomous Navigation of Aerial Vehicle

Guidance of aerial systems is generally based upon inertial navigation sensors and Global Positioning System. Inertial navigation sensors are assumed accurate and reliable for short duration of flights. Error of inertial navigation sensors accumulates and ultimately drifts from actual position increases with flight duration. On the other hand, Global Positioning System (GPS) is independent of flight durations. So errors of inertial sensors are corrected by information of GPS. There is another way to correct errors of inertial navigation sensors. This is based upon image matching techniques. Waypoints are defined along flight path of aerial vehicle. These way points are stored in aerial system as satellite images. When aerial vehicle reaches pre-stored area, it captures image of small geographic area (aerial image). Now image matching algorithm locates aerial image in pre-stored satellite image. Design of these algorithms is a challenging task. Image matching algorithms are computationally very expansive. A number of matching algorithms are evaluated and finally a computationally efficient algorithm is proposed.