Automation anywhere , Scientists are now using drones and machine intelligence to locate meteorites that have just fallen considerably more effectively.
Meteorites provide enticing hints about the early solar system’s characteristics. Finding them, though, is not particularly difficult.
Researchers frequently just spread out across a terrain and wander aimlessly for hours.
Meteorite-hunting –Automation anywhere
According to Seamus Anderson, a planetary scientist at Curtin University in Perth, Australia, a team of six individuals may explore around 200,000 square meters every day while on a meteorite-hunting excursion.
However, searching can take some time because the region over which a cluster of meteorites falls normally cannot be localized to better than a few million square meters, he explains. It goes really slowly.
Search Meteorites by Drone
In 2016, Anderson experimented with the idea of employing drones to photograph the ground in search of meteorites.
That notion developed into a PhD thesis. He and his coworkers published the first successful recovery of a meteorite discovered by a drone in 2022.
The team announced their discovery on August 17 in Los Angeles at a meeting of the Meteoritical Society that they had subsequently discovered four other meteorites at a different location.
Drone-based searches are much faster than the standard way of doing things, Anderson says.
“You’re going from about 300 days of human effort down to about a dozen or so.” It’s also fun and exciting work, he says, but there are challenges too.
Using Drone for Automation anywhere
Drones have been employed by Anderson and his colleagues to look for meteorites in isolated areas of South and Western Australia.
Networks of ground-based cameras that monitor meteoroids flashing through the Earth’s atmosphere inform the team about a fall site. The hunt then begins.
The researchers load a four-wheel drive vehicle with computer and drone gear, battery charging stations, generators, petrol, food, camping gear, tables, seats, and other supplies.
According to Anderson, the trip to the fall location might take more than a day and is frequently made on rocky or nonexistent roads. “You’d better hope you don’t pop a tire.”
The crew then launches its main drone at a height of roughly 20 meters after arriving.
Every 40 minutes or so, when the drone lands to obtain new batteries, the researchers download the data from the drone’s camera, which takes a picture of the ground once every second.
Over 10,000 photos can be captured during an average flying day, which are then digitally separated into almost 100 million smaller parts.
These “tiles,” which are each 2 meters in length, are sent into a machine learning algorithm that has been taught to identify meteorites using pictures of the genuine object or rocks from the actual world that have been painted black.
According to Anderson, the latter are credible substitutes for actual meteorites.
The algorithm is effective but not flawless. When a tile doesn’t have any objects like meteorites, it immediately discards the majority of them – generally more than 99 percent.
However, Anderson estimates that after a day’s worth of flights, there are still over 50,000 tiles that need to be manually examined.
The majority of the time, those tiles include objects that are unmistakably not meteorites, such as animal waste, tin cans, snakes, or sleeping kangaroos.
According to Anderson, those items are labeled as prospective meteorites merely because the algorithm is unfamiliar with them.
It is the team’s responsibility to weed out these false positives.
The researchers send out a smaller drone that flies considerably lower, approximately a meter off the ground, to study items that nevertheless appear plausible to the human eye. Finally, the team visits potential prospects in person.
To prevent objects like excrement and kangaroos from being mistaken for meteorites, the researchers intend to train their system. Additionally, the group is attempting to make its software open-source so that other researchers can utilize it without restriction.
In Antarctica, a center of meteorite sleuthing, Anderson also hopes to see drones in action (SN: 1/26/22).
However, Anderson claims that working in such a remote location will create a whole new set of difficulties, such as ensuring that delicate electronic equipment holds up well in the freezing circumstances. Antarctica is an entirely distinct species.