Visual Navigation Technology In The Application Of The Unmanned Helicopter Landing

Visual navigation is developed from computer vision, which is emerging as a new navigation technology. As a result of the vision navigation has the advantages of strong anti-interference ability, low consumption and low price; it has become the focus of research at home and abroad.In this paper, with the camera projection model, the measurement methods of the relative position and orientation mainly are studied between an unmanned helicopter and the landing pattern.The main content of this thesis is summarized as follows.1. Research on the imaging model of the camera. According to the theory of perspective projection, the pinhole imaging model of a single-purpose is selected. And we research on the coordinate systems definition, the conversion relationship of the position offset and the angle rotation between the coordinate systems.2. Research on the processing methods of the bottom of the image. We analyze the effects on noise processing by image smoothing filter and the methods of feature extraction. And in accordance with the design requirements of the landing pattern, the characteristic pattern is designed. Finally, each corner point is marked with the number after the feature point extraction.3. Research on the position and orientation estimation algorithms based on the corner features and the linear features of the landing pattern by the principle of perspective projection. When the physical coordinates of the four corner points in the characteristic pattern are known, combined with the image coordinates of the corner points,we can calculate the position and orientation between the unmanned helicopter and the landing characteristic pattern. The algorithm based on the straight-line is to estimate the relative position and orientation of the unmanned helicopter by the principle of vanishing points and vanishing lines.The two algorithms are verified in laboratory environment. And the positioning accuracy error is analyzed.The experimental results show that, the estimated accuracy of the position and orientation meets the requirements of the unmanned helicopter landing.