NASA and IBM Release Free AI Tool for Mapping the Moon

NASA and IBM have open-sourced the Lunar Foundation Model, a free AI tool for mapping the Moon. This model is trained on millions of images from NASA's Lunar Reconnaissance Orbiter and other missions, allowing it to easily spot craters, volcanic features, and potential ice deposits.

The Lunar Foundation Model is a type of foundation model, which works differently from traditional AI tools by learning broad patterns from huge datasets. Researchers can fine-tune the base knowledge for specific jobs using a small amount of labeled data, saving time, computing power, and money.

The model has been trained on almost 2 million images, including high-resolution camera shots and multispectral images from NASA's Lunar Reconnaissance Orbiter, as well as data from other missions. It can help find ice deposits, which could supply drinking water, oxygen, and rocket fuel for future missions.

One of the biggest strengths of this model is finding ice-rich zones near the lunar poles, where cold temperatures trap ice for billions of years. The tool also helps speed up crater mapping, a task normally done by hand, allowing scientists to study the Moon's age and history.