Vision-based Navigation For Auto Landing Of The UAV At Night

Machine Vision landing on the autonomy issue is a hot research topic.Visual navigation equipments are simple and inexpensive,and they also have the advantages of complete autonomy and passivity.It has great significance on reducing navigation systems costs and improving the initiative for the unmanned air vehicle(UAV).The research in this paper roots from the program supported by the National Nature Science Foundation.The main work is to obtain navigation information for the auto landing of the UAV during the course of the landing of the UAV by the camera fixed on the UAVExtracting the navigation information for the auto landing of the UAV at night is the principal work in the paper.At night,visual reference is provided by the approach and runway lighting. The pilot lamps of the runway can reflect the location and orientation of the runway,and these lamps which are on the same side of the runway are almost on the same line. The video image processing is done with the image pretreatment algorithm to pop the lights out in the image and reduce the noise at first,the linear features of the lamps on the runway is extracted by Hough transformation in succession,then A sample method for estimating the attitude and position of the UAV is proposed after analyzing the relationship between the UAV's attitude and the shape and position of the lines in the image.It's the core work of this paper and the innovation.Experiment shows the validity of the proposed method.The attitude and position of the UAV are very important to safe landing,these parameters are gained only by vision in this paper.It reduces the dependence on the other equipments and enforces the independence of the UAV.