Vision aided unmanned aerial vehicle autonomy: An overview

摘要

Camera on-board Unmanned Aerial Vehicle (UAV) is not just a device to capture images and video. Computer vision helps UAV senses and recovers environmental information around it. In addition, computer vision techniques, such as stereo vision, optical flow fields etc. play very important roles in UAV autonomy. This paper investigates researches on computer vision applied in UAV autonomy. Visual servoing technique integrates visual methods and UAV flight control. Researches on this topic are classified into three catalogues: vision-based navigation, aerial surveillance and airborne visual Simultaneous Localization and Mapping (SLAM).