Drones, Artificial Intelligence Successfully Recovered Fallen Meteorite in Australia For the First Time

Meteorite Recently Crashed in Australia

Australia's Desert Fireball Network (DFN) is a system of cameras that monitors the sky for fireballs, ensuring that each camera captures every region of the sky so that the trajectory of the incoming space rock can be accurately triangulated, according to Discovery Magazine.

When a meteor enters the Earth's atmosphere, the DFN would immediately detect it and immediately work out the location of the meteorite landing site within a few square miles. Then they will send a team of astronomers to look for the fallen space rock. DFN is considered the first piece of solving the puzzle of finding the meteorite.

On April 1, 2021, the two DFN observatories, one at Mundrabilla station and another at Mundrabilla station, detected a meteor entering the atmosphere for 3.1 seconds. But the two observatories were unfortunately too far from the landing site of the meteorite at 92.5 miles (149 kilometers) and 293 miles (471 kilometers), respectively, Universe Today reported

Due to this, astronomers were uncertain of the exact landing area, although they tried limiting their search within 3.1 miles squared (5.1 kilometers squared), which is still a lot of area to cover. Anderson said that meteorites were special because they are a geologic sample from a region in the Solar System where they came from.

How Did Drones and AI Help in Recovering the Meteorite?

Mr. Anderson explained that usual meteorite searches involved a group of astronomers walking over a large predicted impact site five to ten meters apart to scour the area for more than four days until they find the space rock, Phys.org reported. But that method has a low success rate of approximately 20% since it rarely needed only one trip.

So the team at Curtin University used drones and AI to help search meteorites. They trained the AI to look for fallen meteorites using pictures of known space rocks in similar environments, so they know what to look for before releasing a drone to collect images to feed the algorithm.

After three days, they were able to spot the exact location of the meteorite, thanks in part to the relatively open outback of Australia. Mr. Anderson emphasized that although the algorithm was previously trained on data collected from previous meteorites searches, it still brought them the exact location of the meteorite they were looking for.

He added that novel solutions such as this help make investments in space science and the study of space rocks more cost-effective. Also, it can be used in other applications, like wildlife management and conservation, as the AI can be easily trained to detect objects other than meteorites.

Check out more news and information on Meteorite in Science Times.