social network

Astronomers have found 6,000 new craters on the moon with the help of artificial intelligence

Moon. NASA photo
Moon. NASA photo

Researchers from the Canadian University of Toronto have developed a new way to search for craters on celestial bodies. They trained for this artificial intelligence and found six thousand craters on the moon, which people missed. Scientists have told about this on the university’s website.

One of the developers of the technology Mohamad Ali-Dib (Mohamad Ali-Dib) noted that when it comes to craters, usually scientists manually search, mark and count their sizes in the pictures. This takes a lot of time and resources, so the University of Toronto scientists decided to automate the process with the help of artificial intelligence.

 They used the same technologies on which self-controlled machines operate. Scientists trained artificial intelligence on an array of data with already known craters on most of the lunar surface and tested the neural network on the remaining area. The car found twice as many craters as a man with a manual calculation.
An example of a neural network for finding craters. Left – the result of the algorithm: red circles – new craters detected by the machine, and blue – detected by people manually. On the right, the craters that are missed by the neural network are highlighted in purple.

According to Ali-Dib, researchers for the first time managed to create an algorithm that can recognize craters with high accuracy not only on the Moon, but also on other celestial bodies, for example, on the planet Mercury.

Crater research is an important part of astronomy: having learned the size and location of some craters, scientists can find out new data on the early stages of the formation of the solar system. The moon is convenient for researchers, because due to the lack of atmosphere on it, even the tiniest craters remain for a long time.

Canadian scientists plan to improve the system so that it can accurately recognize craters on other bodies of the solar system. At the time of writing the notes, their research is on a review in the scientific journal Icarus.

Back to top button