Add to Chrome
Add to Firefox
Add to EdgeAutomatic Data Processing for Space Robotics Machine Learning
Oct 03, 2023
Autonomous terrain classification is an important problem in planetary navigation, whether the goal is to identify scientific sites of interest or to traverse treacherous areas safely. Past Martian rovers have relied on human operators to manually identify a navigable path from transmitted imagery. Our goals on Mars in the next few decades will eventually require rovers that can autonomously move farther, faster, and through more dangerous landscapes--demonstrating a need for improved terrain classification for traversability. Autonomous navigation through extreme environments will enable the search for water on the Moon and Mars as well as preparations for human habitats. Advancements in machine learning techniques have demonstrated potential to improve terrain classification capabilities for ground vehicles on Earth. However, classification results for space applications are limited by the availability of training data suitable for supervised learning methods. This paper contributes an open source automatic data processing pipeline that uses camera geometry to co-locate Curiosity and Perseverance Mastcam image products with Mars overhead maps via ray projection over a terrain model. In future work, this automated data processing pipeline will be leveraged for development of machine learning methods for terrain classification.
