The adaptive compensation algorithm for small UAV image stabilization

Today, there are an increasing number of Unmanned Aerial Vehicle (UAV) platforms being equipped with video for real-time observation. Unmanned Aerial Vehicle Systems (UAVs) have been generally used in a great variety of missions because of its inexpensiveness and flexibility. However, UAV often experience unintended translation and rotation due to atmospheric turbulence. And it leads the compensation image to losing lots of information when the large drift occurs. In order to get rid of the unwanted jitter and retain the most information, an estimate of the intentional motion is necessary. This paper presents a novel approach to estimating the intentional motion by adaptive compensation algorithm. And it can retain image information and remove image jitter when the pre-path is deemed to be a straight line. The method has been demonstrated through a series of experiments on real aviation video data.