A General Model and Application Analysis of Spatial Resolution on Large Space Navigation Imaging of Yutu 2 Lunar Rover

The Yutu 2 lunar rover has been engaged in scientific research on the lunar surface for more than three years. In order to maximize the efficiency of planning and driving for the Yutu 2, the distance between the beginning and end stations of each movement is large. This causes significant scale differences and complex affine transformations between the images in adjacent stations, which usually cause failure of image matching and navigation. To address the above problems, current methods use the single-dimensional pixel spatial resolution model based on imaging geometry to analyze the reasons for failed or successful image matching. However, they are not adaptable to images with complex affine transformations. In this article, a general 2-D pixel spatial resolution model based on the algebraic derivation of the homography transformation is proposed. With the proposed model, the scale difference between images of adjacent stations can be quantitatively analyzed, and the rate of successful matching between images that have complex affine transformation and scale differences is increased. This will benefit engineering applications, such as visual localization and navigation. Experiments under various conditions, such as the horizontal and inclined lunar surface, are performed to verify the effectiveness of the proposed method.