The Application Of Visual Navigation In Auto-landing Of Unmanned Aerial Vehicle

Computer vision based navigation is a high accuracy and adaptive method, so it has attained more and more attention. Computer vision technology was applied to auto-landing navigation of unmanned aerial vehicle (UAV) by using image tracking technology in this paper.Firstly, the main navigation method is discussed to get their pros and cons. Then a visual navigation method is described which is based on the transit's control system and the image tracking technology. After discussing the scheme of that method, integrated navigation is studied using INS.Secondly, the image tracking methods are studied mainly in this chapter. Feature extraction algorithms have been considered. Mainly the edge-invariant-moments algorithm and gradient-angular histogram of image edges have been described. And for the target tracking, the corner features of images have also been extracted. After that, the image tracking technology is discussed, such as the edge tracking algorithm and the moving object detection by using background subtraction technology. The results of the experiment demonstrate that the algorithms used for navigation and image tracking offer a very good real-time performance.Camera calibration is introduced in the application of the visual navigation used during the auto-landing of the UAV, which can influence the positioning accuracy. Then the relationship of the angles of the transit and the image coordinates is shown to apply to the transit control, and get the good result in the experiment. Finally, primary content of this task has been simply introduced and research aspect for future work has been highlighted..