High precision localization of the Chang'E-3 lunar rover

High-precision localization technology is an important component for conducting scientific explorations in the Chang'E-3 lunar rover mission. In this study, the exact stereo images strips network localization (SISNL) method is proposed to solve the localization problem of the lack of high-precision control points on the lunar surface. First, we obtained the camera calibration parameters using the bundle adjustment algorithm in a ground field with some known control points. Hence the baseline and relative orientation elements of the stereo cameras were acquired. Second, after building the stereoscopic model using feature matching and forward intersection, the stereo images strips network was generated, using the sequence of connection points at different stations. Third, based on the weighting matrix of the coordinate observations of the connection points, and on the extended orthogonal Procrustes theory, the least squares adjustment algorithm was used to solve the mathematical model of the images strips network. The lunar rover's position and attitude information was then acquired directly. At the end of the study, we compared the SISNL method with the traditional bundle adjustment method using the indoor testing field data. Though the two methods have equivalent localization accuracy, the comparision results show that the SISNL method was superior to the latter in terms of both stability and efficiency. On this basis, the SISNL method was successfully applied to the Chang'E-3 lunar rover teleoperation system and satisfied the demand for lunar rover localization.