NASA and IBM Release Open-Source AI Model to Map the Lunar Surface
NASA and IBM have shared an AI model trained on nearly two million spacecraft images to map the lunar surface, locate hidden water resources, and study volcanic features.
NASA and IBM have launched a joint open-source foundational AI model trained on nearly two million spacecraft images to study the lunar surface and detect hidden water spots.
AI-Powered Lunar Mapping
While planetary scientists used to spend hours examining satellite images, an AI model developed through a partnership between NASA and IBM now processes surface data at unprecedented speeds.
Hidden Water Resources in Polar Regions
As temperatures in unilluminated craters at the Moon's poles drop below minus 250 degrees Fahrenheit, the developed model helps detect ice deposits in these areas.
Training Data and Performance Tests
While the AI learns from one million high-resolution photographs at a three-foot pixel scale along with thermal and mineral maps, it reduced the error rate by 22 percent in benchmark tests.
Future Research and Drilling Operations
Although the system flags potential ice areas, it cannot see directly beneath the soil, meaning next-generation drilling rovers will need to physically verify the water deposits.
Volcanic Features and Solar System History
The model successfully mapped rare volcanic features known as irregular mare patches, indicating that the Moon may have experienced volcanic eruptions in the relatively recent past.