In this paper, the method vision-based for UAV autonomous landing is researched using technology of image processing, computer vision and projection geometry. The first task of computer vision-based UAV autonomous landing is to recognize and track the runway using computer vision. A simple method is discussed to recognize and track the runway in the image sequences using template matching. Secondly, an algorithm of image processing is introduced which can detect horizon of airfield and edges of runway commendably. And then, based on work of former researchers, it is developed which is a software system of UAV's autonomous landing image processing and parameter-obtained. A corresponding simulation experiment has been done making use of this software. Lastly, an algorithm is offered to obtain UAV's flight parameters based knowledge of projection geometry. One strongpoint of this algorithm, the yaw of UAV can be gained expediently if the roll of UAV is known. This algorithm also has been proved by corresponding simulation experiment.Author discusses elementary methods vision-based for UAV autonomous landing in this paper. All this is useful for the realization of UAV autonomous landing.
|