Research On Navigation Algorithm Based On Binocular Vision For A Lunar Rover

A lunar rover is the most direct and effective vehicle for detailed lunar exploration of a wide range,and the autonomous navigation technology is the key to implement the fully independent exploration for a lunar rover. Therefore, methods of autonomous navigation are considered as critical problems in lunar rover technology. This dissertation studies a kind of autonomous navigation method based on binocular vision for a lunar rover.Firstly, a navigation algorithm flow chart is given through the analysis of the navigation principle based on binocular vision, and correlative coordinate systems and their transformational relations are defined. The imaging model of camera, as well as the ranging model of binocular camera is constructed. And then, the method of acquiring the coordinates of points in space based on binocular cameras is derived.Secondly, the dissertation confirm the Susan algorithm as the corner detection method in the visual navigation system of a lunar rover, through analyzing Harris corner detection algorithm and Susan corner detection algorithm and taking a comparison between them. Then, a gray correlation stereo matching algorithm and a corner tracking algorithm based on uniqueness constraint and ordering coherence constraint are derived, verified by simulation experiments.Thirdly, the quaternion is chosen to describe the motion of a lunar rover according to a comparison of several rigid motion description methods. Besides, a navigation motion parameter precise estimation algorithm based on the unit quaternion method is derived, verified by mathematical simulation experiments.Finally, a improved Ant Colony Algorithm is derived in order to plan path for a lunar rover, which solves the limitations of slow convergence and easy falling into local optimal solution in the traditional algorithm by using rasterization, bidirectional searching, improved pheromone updating strategy and improved spread methods. The algorithm improves path planning ability of a lunar rover in the navigation.