AI solutions for space rock research

  • 3 minutes

If we integrate AI into the process of analyzing space rocks, we can improve the collection process for both humans and rovers. We could send astronauts to the Moon armed with a computer that can take photos of rocks. The computer could show the astronaut the rock type. The astronaut could determine if that type of rock is needed in the collection and decide to pick it up or leave it. In a future mission, the computer could be placed in a rover that autonomously drives across the surface of the Moon and scans for rocks that we need for research.

By integrating AI, astronauts could more quickly and accurately locate and identify rocks they should bring back to Earth. The computer could collect metadata like location, temperature, and light exposure. With accuracy feedback from astronauts and scientists on Earth, the AI model used to identify valuable rocks would improve over time.

How can AI help research on Earth?

As we mentioned earlier, AI works better when the sample data is cleaner. We see improved performance for multiple scenarios involving space rocks:

  • When the rocks are physically clean
  • When the photos of the rocks have similar lighting, so the visual coloring of the rock is consistent
  • When the edges between the rock and the background are clean and consistent
  • When each photo of a rock has a clear indication of scale

An AI model that can help identify, classify, and track these rock samples would provide clear instructions about how to take the photos to fit the current model. When we have an AI model that fits these criteria, lunar rock curation and research can move toward answering even more nuanced research questions. We might even consider using a rover and AI on Earth to collect meteorites in Antarctica.

How can AI help with research in space?

AI and image quality might not be good enough yet to take a satellite picture of a planet's surface and give an overview of the types of rocks that are on it. But, we can start to understand where AI can help and where it's limited.