Lunar Anomaly Detection: How AI is Transforming Moon Exploration

Artificial intelligence is transforming the way we explore the Moon. In a recent SETI Live, authors Adam Lesnikowski and Daniel Angerhausen, hosted by Franck Marchis, a Senior Planetary Astronomer at the SETI Institute, discussed how AI is revolutionizing planetary science and helping us detect unusual features on the Moon.
 

The Lunar Reconnaissance Orbiter

NASA’s Lunar Reconnaissance Orbiter (LRO), launched in 2009, has been instrumental in mapping the Moon’s surface. Over the past decade, it has captured thousands of high-resolution images—far more than humans can manually review.

To tackle this challenge, scientists have developed an automated system that quickly identifies scientifically significant images from the LRO data, making it the first anomaly detector for planetary imagery. The AI is trained on millions of images and reconstructs them from compressed information.
 

AI’s Role In Lunar Exploration

Here are some of the key advantages of this approach: 

1. The use of unsupervised learning reduces human bias.

2. Unlike traditional methods, this AI detects anything that deviates significantly from the norm, allowing it to discover unexpected lunar features that might have otherwise gone unnoticed.

The AI has already detected several significant anomalies on the moon:

• Volcanic pits & skylights: Collapsed sections of lava tubes that could serve as future lunar habitats, offering protection from radiation and temperature extremes.

• Fresh impact craters: Help scientists study meteorite impacts and assess risks for future missions.

• Apollo landing sites & landed assets: The AI successfully identified human-made structures, proving its ability to detect technosignatures.

• Irregular mare patches: AI has flagged these mysterious geological features for further investigation.

Expanding AI’s Capabilities Beyond the Moon

While the algorithm has effectively spotted known anomalies, the next goal is to expand its reach to the entire lunar surface, requiring additional funding and computational power. The team is also exploring commercial applications for lunar prospecting, which could aid in resource identification and extraction efforts. 

Beyond the Moon, this AI-driven approach could be applied to celestial bodies such as Mars, Mercury, and Ceres. Additionally, because the algorithm is data-agnostic, it could assist in other scientific fields, including analyzing light curves from space telescopes like NASA’s Kepler and TESS.
 

Challenges in AI-Driven Discovery

Despite its promise, AI-driven anomaly detection still faces challenges. Potential biases in AI models and the need for large, high-quality datasets must be addressed to ensure accurate findings. The team has made its data and code publicly available to encourage further research and validation.

With AI unlocking new ways to detect lunar anomalies, the future of planetary exploration is taking an exciting turn. Watch the video below for a detailed presentation on automated lunar anomaly detection or check out “Looking for Lunar Anomalies Using Automated Methods” on our YouTube channel.