A new AI algorithm finds a “potentially dangerous” asteroid near Earth

“By demonstrating the real-world effectiveness of the software Rubin will use to search for thousands of unknown, potentially hazardous asteroids, the discovery of 2022 SF289 makes us all feel even more confident,” Rubin said in a statement. Ari Hines, HelioLinc3D lead developer and researcher at the University of Washington.

Scientists search for potentially hazardous asteroids (PHAs). telescope systems Such as the ATLAS survey led by a team from the Institute of Astronomy at the University of Hawaii. They do this by taking pictures of parts of the sky at least four times each night. The detection is made when they notice a point of light moving unmistakably in a straight line over the series of images. The researchers found out about 2350 using this method, but they estimate that there are at least a few more waiting to be discovered.

From its peak in the Chilean Andes, the Vera C. Rubin Observatory will join the search for these objects in early 2025. Rubin’s observations will greatly augment discovery rate. Rubin will scan the sky at unprecedented speed with his 8.4-meter-diameter mirror and massive 3,200-megapixel camera, visiting points in the sky twice a night instead of the four times current telescopes require. But with this new rate of observation, researchers need nA new kind of algorithm Detection to reliably detect space rocks.

Rubin’s Solar System Programs team at the University of Washington’s DiRAC Institute is conducting research to develop such codes. Together with the Smithsonian Institution’s Chief Astrophysicist, Matthew Holmanwho in 2018 pioneered a new class of heliocentric asteroid search algorithms, researcher Show Heinz And Siegfried EgelA former University of Washington researcher has developed HelioLinc3D: a code that can find asteroids in the Robin data set.

Discover a new asteroid

With Rubin still under construction, Heinze and Eggl wanted to test HelioLinc3D to see if it could detect a new asteroid in the existing data, with too few observations to be detected by the current conventional algorithms. John Tunney And Larry DinoAtlas astronomers submitted their data for testing. Rubin’s team configured HelioLinc3D to search this data and on July 18, 2023 discovered the first PHA: 2022 SF289, imaged by ATLAS on September 19, 2022, at a distance 20 million km from Earth.

In hindsight, ATLAS 2022 observed SF289 three times on four separate nights, but four times in one night were not identified as New relative object to ground. But these are the times when HelioLinc3D excels: it successfully combined bits of data from all four nights and made the discovery. “Any study would have difficulty detecting objects like 2022 SF289, which are close to their sensitivity limit, but HelioLinc3D shows that it is possible to recover these faint objects as long as they are visible for several nights,” Denneau explained. “This, in effect, gives us a bigger and better telescope.”