Visual navigation algorithm based on line geomorphic feature matching for Mars landing

Wei Shao

doi:10.1016/j.actaastro.2020.04.055

2020-08-01

PRE

DOI：10.1016/j.actaastro.2020.04.055

原文链接

原文求助

摘要

En 中文

In order to satisfy the requirement of precise pin-point landing mission in Mars exploration, this paper proposes an innovative visual navigation algorithm based on feature lines extraction and matching to estimate the attitude and position of the lander during descent phase. Currently, common navigation landmarks are feature points and craters. But, the number of feature points extracted from landing image is so large that the image process is complex and time-consuming, and they do not usually have absolute coordinate information to absolute navigation. Meanwhile, crater is matched difficultly. Hence, ridges, fissures and other surface textures similar to lines are most common visual features of Mars and can be used as navigation landmarks for visual absolute and relative navigation. Firstly, that feature lines matching based on the EDLine algorithm remains invariant to scaling, rotation, and illumination, etc. provides the basis for visual navigation by using feature line. Secondly, feature lines selected as navigation landmarks can be used to absolute navigation and relative navigation respectively, which can implement integrated navigation. In this paper, a novel integrated visual navigation has been designed to finish the precise pin-point landing and overcome the accumulated errors of relative navigation and the limited application range of absolute navigation. Experimental simulation results show that the integrated navigation algorithm solves the limitation of single navigation mode and improves the effectiveness and stability of the algorithm.

Keyword:

Mars landing

Visual navigation

Pin-point landing

Feature lines matching

Integrated navigation