We have spent centuries documenting our solar system and galaxies. The biggest problem with mapping space is it’s infinite size, which directly mean it will take long time to look at it all. This is why artificial intelligence has been such a boon to this science.
Recently a new computer algorithm, artificial intelligence helping to draw a more detailed map of the moon. This new AI is based on lunar mapping technology. It has nearly double our index of known moon craters within a few hours.
The new lunar mapping technique successfully counted new pockmarks on the moon. The technology inadvertently discovered nearly 7,000 new craters, through available datasets from previous lunar observation information.
Moon is dotted with a vast number of craters. Traditionally, we’ve counted and observed this craters by visual inspection, which a very time consuming and limited process. In this method we can only document what we can see. But by use of AI we have found an another method of discovering our space that also within a short span of time.
The project is led by Ari Silburt at Penn State University and Mohamad Ali-Dib at the University of Toronto. The aim of the new crater-counting AI is to increase the accuracy of moon crater documentation. The study is published in the journal Icarus.
According to New Scientist, the tool used for this particular research is what’s known as a convolutional neural network or CNN. CNN’s are particularly good at sorting through visual data.
Through a dataset covering two-thirds of the moon, the AI studied lunar images to learn what craters look like. It was taught an algorithm designed to identify craters larger than 3 miles (5 kilometers) in diameter. This limit was set to help the AI distinguish between a crater and a planetary feature, such as a mountain edge.
The algorithm quick gain it’s accuracy, it not only start to document already known craters on the moon’s surface but also identify previously unknown ones. Although the technology falsely identified a few non-craters, in only a few hours the AI was able to nearly double the number of known craters on the moon measuring 3 miles or more by finding 6,883 more.
The AI wasn’t perfect, but result were satisfactory. The new craters discovered, 15 per cent are smaller in diameter than the minimum crater size in the ground-truth dataset. The errors compared to the human-generated datasets are only 11 per cent or less, making the deep-learning tool useful in automatically extracting crater information on various solar system bodies. The same network also successfully detected craters on Mercury, which has a completely distinct surface compared to the moon, the report said.
The scientists who conducted the work, from the universities of Toronto, Penn State, and Arizona State, write in their paper, the system was consistent and, most importantly, fast. “Once trained, our CNN greatly increases the speed of crater identification, taking minutes to generate predictions for tens of thousands of Lunar DEMs,” they write. (A DEM is a digital elevation map, and it is the standard image-type used to find and classify craters.) “This is, of course, all done passively, freeing the scientist to do other tasks.”
This program could be used to catalog impact scars on other moons or planets, which might improve scientists’ understanding of how various objects roamed our solar system in the past. It can be used for spotting gravitational lenses, discovering new exoplanets, identifying pulsar stars, and classifying galaxies.
More in AI :