Study On UAV Positioning Algorithms Based On Image Tracking

Image navigation is the technology which gets navigational and location parameters by analysing and handling the image acquired from sensors carried by plane. Image navigation has been the focus of both home and abroad in the navigational field because of its advantages including abundant information, low cost and independence.Based on various common navigational methods, this paper deeply analyses the image navigational methods of UAV field, including the research status, algorithmic principles and characteristics. The most matured research result is to get the information of homonymous image points of multiple ground targets by a real-time image, and then solve the parameters of position and posture of UAV by camera and measure theory.This method requires the navigational computer could store a amount of matching maps, and the number of homonymous points on real-time image should be three or more. Aiming at above problems, this paper proposes a single point orientation algorithm, based on image track. If the distance from UAV to ground target can be measured by laser telemeter, we can get the precise UAV real-time location parameters by active orientation algorithm. However, laser telemeter has some disadvantages, including being exposed and interfered easily.Aiming at the limitation of laser telemeter,a passive single point orientation algorithm is proposed in this paper. The algorithm follows the trail of ground target, then makes use of informations including the coordinates of real-time image points, platform angles of optical axises and posture angels of UAV to solve the UAV location parameters according to the relationship of space geometry. The scope of image navigation of UAV can be expanded by combining active single point orientation algorithm and passive single point orientation algorithm. Besides, the single point orientation algorithm based on image track is deduced in this paper, the simulation results shows that the algorithm is effectual.Finally, the single point orientation algorithms proposed in this paper and the Strapdown Inertial Navigation System are united to constitute a navigation system. This paper focuses on the deduction of Kalman filtering state equations and the measurement equations of the system.