The Application Of Computer Vision In Lunar Rover Navigation

The moon, nearest to the earth and best known to us, can be used as a perfect transferable station in the space research in future and a base for human beings to enter the solar system. At present, the exploration to the moon could provide a better support in science and technology for a deep space exploration in future. After an elementary study with lunar orbiting satellite, it is necessary to make exploration devices land on the moon and carry out detailed exploration with lunar rover in a rather large area. These efforts will lay a foundation for landing man on the moon and establishing lunar base.For the unmanned lunar rover work safely in a long enough time in complex environment of lunar surface, a good navigation system is needed to supply position and pose message in real time to support automatic navigation and remote control. To synthesis the advanced navigation technology and the characteristic of the lunar rover, the computer vision is selected as the main mode of the navigation system of the lunar rover.In this paper an algorithm of monocular vision navigation is designed. The implementation uses two-dimension images as input. The arrangement of light spots on the lunar rover and the imaging geometry of the camera are used to compute the three-dimension coordinates of the light spots in camera coordinate system, and after transformation, the world coordinate of the light points. We can then easily get the coordinate value of center of mass of the rover after averaging the world coordinate value of light points. At last the relatively attitude can be measured by the transformation between vehicle body coordinate system and the world coordinate system of N light points. The algorithm can be confirmed by the result of simulation. For the reason that this method use the arrangement of characteristic light spot directly to avoid the aberrance that come out during calculating in traditional method, it can improve the estimation precision of position and attitude. Besides, it provides a certain degree of redundancy to the visual navigation system.