Vision-based Estimation Of UAV Attitude

UAVs have a very board application prospects in both military and civilian fields because of their advantages such as small, low-cost and high security. Traditional navigation cannot ensure the precision since the inertial navigation has cumulative error and the GPS navigation may be useless when the GPS signal is missing. Vision-based navigation solves the navigation parameters using the characteristic information derived from image processing. It is autonomous and low-cost. So the vision-based navigation as aided navigation way becomes a research focus.The main content of this paper is the attitude estimation method. The main works are as follows:(1)Extensive explorations have been conducted about the basic filtering, edge detection, image segmentation and other image processing methods, the navigation system coordinate system conversion and the relative relationship between the camera coordinate system in computer vision, image coordinate system and world coordinate system. This section provides a theoretical basis for follow-up work.(2)An effective horizon detection method was proposed. Firstly, separate the image using method based on OSTU method before conducting the edge detection. Then derive the horizon line using Radon transform or Hough transform when the edge is detected. Compared with the method based on edge detection, the proposed method has less misjudgment. This has demonstrated the effectiveness of the proposed algorithm.(3)An attitude estimation method based on the horizon line was proposed. Establish the relationship between the horizon and the attitude angle according to the linear camera imaging projection model. The attitude angle of UAV can be calculated by the slope and intercept of the line equation.(4)Experiments were designed and conducted to verify the proposed method after the camera calibration has been done. The results show that the proposed method is feasible and correct.