Abstract
Background In recent years, we have seen renewed efforts to study and explore the Moon; modern techniques like machine learning can be important in this context, especially in recognising and classifying the lunar surface. The EXPLORE Machine Learning Lunar Data Challenge was a public initiative during the last quarter of 2022. Its objective was to encourage participants to apply machine learning techniques to identify potential hazards for a planetary mission and to design a robotic traverse for exploring the lunar surface. Methods The lunar region targeted by the challenge was the Archytas Dome in Mare Frigoris, a location with a varied geology and a potential zone for future exploration. We provided training datasets of craters and boulders to the participants, who used them to complete the three steps of the challenge: creating a model that detects these landforms, applying these models to the Archytas Dome region, and computing a traverse for optimal exploration of the zone. In this paper, we showcase the results and considerations of the team that won the challenge. The first step was to enhance the training data by generating new labels and resizing the existing ones. The original and the improved dataset were then used to train four iterations of a neural network model. Results The model with the enhanced dataset yielded the best scores when applied to the Archytas Domes zone (75.46\%). Finally, the traverse was calculated using proximity analysis while avoiding steep slopes and dangerous landforms. Conclusions We found that the variations between tasks and the different approaches necessary to solve them turned out to be the major difficulty of the challenge, as it required backgrounds in both remote sensing and computer sciences. This was reflected in the low participation and the multidisciplinary of the members of the winning team.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.