视觉里程计
地形
计算机视觉
计算机科学
人工智能
里程计
航天器
遥感
地质学
机器人
移动机器人
地理
航空航天工程
工程类
地图学
作者
John A. Christian,Lillian Hong,Paul McKee,Randall Christensen,Timothy Crain
出处
期刊:Journal of Spacecraft and Rockets
[American Institute of Aeronautics and Astronautics]
日期:2021-01-01
卷期号:58 (1): 164-181
被引量:15
摘要
Image-based terrain relative navigation is expected to play an important role in the safe operation of upcoming lunar exploration missions. This work provides a detailed treatment of how visual odometry direction-of-motion measurements may be constructed using images from a monocular camera and without the need of an onboard map of the lunar surface. Substantial care is required to achieve best-possible navigation performance, which numerical studies indicate is sufficient to enable many types of autonomous navigation. Results are shown for historical Apollo images and for synthetic images. While visual odometry alone may not meet every mission’s needs, it is a powerful technique that should be a part of every professional spacecraft navigator’s toolkit.
科研通智能强力驱动
Strongly Powered by AbleSci AI