AI Discovers More Than 27,000 Overlooked Asteroids in Old Telescope Images

More than 27,000 asteroids in our solar system had gone unnoticed in the images of existing telescopes, but thanks to a new algorithm powered by artificial intelligence, we now have a catalog of them. The scientists behind the discovery say the tool makes it easier to find and track millions of asteroids, including potentially dangerous ones that could collide. Land some day. It is for those menacing space rocks that the world would need years of warning before trying to divert them far from our planet.

Most of the newly discovered asteroids float in the asteroid belt between Mars and Jupiter, where scientists have already cataloged more than 1.3 million such rock fragments in the last 200 years. The latest bounty, discovered in about five weeks, also includes about 150 space rocks whose trajectories take them within Earth’s orbit; However, to be clear, none of these “near-Earth asteroids” appear to be on a collision course with our planet. Others are trojans that follow Jupiter in its orbit around Sun. Observations of these asteroids have yet to be submitted and accepted by the International Astronomical Union’s Minor Planet Center, the official body responsible for asteroid discoveries.

Astronomers conventionally find new asteroids by studying areas of our sky over and over again, through telescope images collected several times each night, usually every few hours. While the planets, stars and galaxies in the background remain unchanged from image to image, asteroids are seen as noticeably moving specks of light, which are then pointed out and verified. From there the orbits of these asteroids are determined and monitored.

Related: How AI Could Help Find Alien Planets and Asteroids

“This is really a job for AI,” Ed Lu, executive director of the Asteroid Institute and co-founder of the B612 Foundation, said early last month during a discussion about the discovery. In fact, artificial intelligence tools designed for asteroid hunting are already approaching levels achievable by humans, Lu said: “I think we’re going to surpass that quickly in the coming weeks.”

The algorithm Lu’s team developed, known as Trackletless Heliocentric Orbit Retrieval, or THOR, analyzed more than 400,000 archival images of the sky maintained by the National Optical-Infrared Astronomy Research Laboratory, or NOIRLab. As long as there are about five observations in 30 days associated with the same area of sky, the algorithm can go to work. It is trained on a large data set that makes it capable of analyzing up to 1.7 billion points of light in a single telescope image. It is designed to observe and connect one point of light in one image of the sky to another in a different image, and determine whether both specks represent the same object; Most of the time, that indicates an asteroid moving through space, according to a description of the algorithm by the B612 Foundation.

“We don’t own a telescope, we don’t operate a telescope,” Lu said during the discussion. “We’re doing this from a data science perspective.”

The scientists expanded their algorithm using Google Cloud, whose computational power and data storage services made it easier for scientists to test thousands of orbits of asteroid candidates, according to a statement released Tuesday (April 30) by the B612 Foundation.

“Not only can we find asteroids in data sets that were never designed to do so, but we can also make all the other telescopes in the world better at finding asteroids,” Lu said during the talk. “It’s a change in the way astronomy is done.”

In 2022, the same team of scientists used THOR to discover 100 asteroids that had not been detected in images from existing telescopes. Other teams of astronomers have also leveraged AI to find new asteroids. Just two weeks ago, for example, citizen scientists led the training of an algorithm that led to the discovery of 1,000 new asteroids in archival images clicked by the Hubble Space Telescope. Last July, software called HelioLinc3D designed to search for near-Earth asteroids found a 600-foot-wide (180-meter-wide) space rock that was expected to come within 140,000 miles (225,000 kilometers) of Earth. That’s closer than the average distance between our planet and the moon.

So far, scientists have detected more than 2,000 “potentially dangerous asteroids” and estimate that about 2,000 more are yet to be discovered. Detecting these space rocks in an effort to aid planetary defense is one of the tasks of the upcoming Vera C. Rubin Observatory in Chile, for which the HelioLinc3D asteroid search software was developed.

The 8.4-meter telescope, which will begin operation next year, will take images of the southern sky every night for at least a decade, with each image covering an area of 40 full moons. Scientists say this cadence, supported by AI-based software like THOR and HelioLinc3D, could help the observatory find up to 2.4 million asteroids (double the number now cataloged) in its first six months of operations.

St John Shopon

Us News Update

AI Discovers More Than 27,000 Overlooked Asteroids in Old Telescope Images