An intended motion estimation method based on unmanned aerial vehicle aviation video image

Abstract

In the process of the UAV aviation video image stabilization, whether the image sequence interval used to estimate the intended motion is appropriate will have effect on the stabilization quality, either too stable or unstable. To solve this problem, a method of intended motion estimation using the UAV's flight parameters as auxiliary data is proposed in this paper, which also helps to realize real-time image stabilization. UAV always has real-time access to UAV position, flight attitude, velocity and other flight parameters. In this paper we analyze the relationship between these parameters with the flight pattern as a basis of the change in intended motion to achieve an effective judge of the estimation sequence, thus obtain accurate parameters of the intended motion. The method has been demonstrated through a series of experiments on real aviation video data.