Research on visual localization method of lunar rover

Scientific investigation using the lunar rover is based on its position. Positioning the lunar rover is a key capability. This study presents a positioning method based on computer vision. The scale-invariant feature transform(SIFT) algorithm, correlation coefficient matching algorithm, least squares matching algorithm and bundle adjustment model are used together to navigate and position the rover between adjacent lunar stations.A stereoscopic navigation system was constructed in the laboratory, and used to test the feasibility and accuracy of the method. The results validated the relative accuracy of visual orientation to better than 4%. The algorithm was successfully applied to the positioning of the Jade Rabbit lunar rover.