Drone autonomy learns to find the ‘hidden’ meteorite impact site

It’s easy to find big meteorites (or their crater) after they reach the earth, but the smaller ones are often ignored – scientists recovered more than 2 percent of them. However, soon, it might be a question to send a robot to do work. Today’s universe reported that researchers have developed systems that have autonomous drones using learning machines to find smaller meteorites at the clash location which is ‘hidden’ (even if observers are traced) or inaccessible.

This technology uses a mixture of convolutional neural networks to recognize meteorites based on image training, both from online images and shots performed from the team collection. This helps AI distinguish between rock and ordinary stone rocks, even with various forms and conditions of the field.

The results are not perfect. While the test drone does the meteorite planted properly, there are also some false positives. Can be a short time before the robot plane is quite trusted to provide accurate results themselves.

The implications for space science are significant if technology proves accurate. This will help scientists see and potentially restore meteorites that are too small to be found or too far away. That, in turn, can help show the source of meteorites and identify the composition of the stone. Simply put, drones can fill the gap in humanity’s understanding of cosmic debris that landed at our door.