Fig. 2: The reconstructed trajectory of Chang’E-6 and landing site represented in different DOMs from LROC.

The upper section presents the visualization and analysis of the reconstructed trajectory, while the lower section shows the Chang’E-6 landing-site localization results, from a global to a local view, along with the corresponding mapping of the results. a The reconstructed trajectory of the frames sampled from the complete descent image sequence. b The computed 3D scene. c Chang’E-6 landing zone captured from LROC quick-map website (https://quickmap.lroc.asu.edu). The rectangles in b, c represent the same area and the red cross is the landing site. d The reconstructed trajectory using the complete descent image sequence. f, h Enlarged views of regions within d. i Enlarged view of the region in h. Chang’E-6 underwent two significant orientation maneuvers during its descent phase. The first maneuver happened above f, the second above h and each time Chang’E-6 became more vertical to the lunar ground. g, e The camera perspective before and after the first maneuver. i indicates that the autonomous landing system on Chang’E-6 adjusted orientation to ensure safe and precise landing. The black arrow indicates that the lander retreated by 23.5 m, while the angle illustrates that the direction of the lander deviated by 134.8 degrees from its original trajectory. j The projected DOM region. It can be observed that Chang’E-6 landed on the relatively flat region. k, l Cropped images derived from LROC DOMs captured before and after Chang’E-6’s landing. The yellow points represent manually selected matching feature points. The red cross marker in k aligns with the reference point in the original imagery, while the red cross in l is positioned through homography matrix projection calculated from the matched points. The spatial relationship between these markers validates the coordinate transformation accuracy.