Release of open AI model trained on 17 years of lunar data maps ice, craters and volcanoes

NASA has unveiled a groundbreaking open-source AI model designed to revolutionize research on the moon's surface. This collaboration with IBM Research and academic institutions enables researchers to analyze vast lunar data collected over 17 years by NASA's Lunar Reconnaissance Orbiter (LRO). The model, named NASA-IBM Lunar Foundation Model, is hosted publicly on Hugging Face, allowing scientists worldwide to utilize its capabilities easily.

The model was trained on over 2 million image tiles from the LRO dataset, which comprises high-resolution camera images at 1-meter resolution and nearly 964,000 multispectral images at 100-meter resolution. It also incorporates data from other lunar missions, such as NASA's GRAIL, Lunar Prospector, and JAXA's Selenological and Engineering Explorer. By leveraging this pre-existing knowledge, scientists can quickly analyze the moon's surface, focus on interpreting findings, and determine implications for future exploration.

One of the model's key strengths lies in its ability to estimate polar ice stability, which is crucial for understanding the moon's history and potential resources for future space missions. Additionally, the NASA-IBM model accelerates the identification and measurement of craters, enabling scientists to reconstruct solar system history more accurately.

In essence, this open-source AI model represents a significant leap forward in lunar research, transforming how scientists approach the study of Earth's only natural satellite.