A team of researchers, with the support of the ETH Zurich , created an artificial intelligence system that studied the moon’s craters, which almost never get sunlight. Scientists have long been interested in such depressions, because large reserves of ice can be hidden at their bottom.
The neural network is based on images taken by the Lunar Reconnaissance Orbiter probe. With the help of machine learning-based AI Hyper-efficient nOise Removal U-net Software (HORUS), researchers were able to make visible dark areas in the photo of 44 dark craters with a diameter of more than 40 meters. The results of this processing were disappointing: at the bottom of the craters there are no significant ice masses that scientists expected to find.
Researchers believe that there is probably ice in such dark corners of the Moon, but not in its pure form, but mixed with lunar soil.
The discovery of ice in craters could have played an important role in the development of the Moon, but it did not work out. The work done has brought only indirect benefits for further research, because now scientists have large amounts of accurate data on the satellite’s topography. Experts are confident that this information will greatly facilitate future expeditions, including for the upcoming Artemis with the landing of astronauts in 2024.
Visible routes to permanently shadowed regions can now be designed, greatly reducing the risk to Artemis astronauts and explorer robots.
This is especially valuable, as the suit will only provide a limited amount of time in the cold of the shadowed craters; modern solutions allow you to stay there for no more than two hours.
Read also 🧐