Abstract: This paper comprehensively utilizes three key data sources to prepare the inputs needed for descent image navigation: Chang’e6 lander position and attitude data, high resolution lunar surface images captured by Chang’e-6’s cameras, and detailed lunar terrain information. Based on Chang’e-6’s descent imagery data, this paper uses orbital data to establish an initial landing zone model, then applies advanced computer vision techniques to register and fuse camera images with terrain information, validating the feasibility of visual navigation through orthophoto feature point matching and crater recognition matching methods. Visual navigation positioning can be fused with inertial navigation positioning data, significantly improving positioning accuracy and reliability. Research results show that visual navigation methods can effectively correct cumulative errors from inertial navigation and serve as a key component of integrated landing navigation. This navigation technology is not only applicable to the Chang’e-6 mission but can also be extended to other lunar and planetary exploration missions, providing strong technical support for future deep space exploration.

Key words: Chang’e-6 imagery, visual navigation method, the image terrain matching;deep space exploration