A group of astronomers and amateur citizen scientists have discovered more than a thousand previously unknown asteroids in 37,000 archival images taken by the Hubble Space Telescope over the past 20 years. These images also helped train a machine-learning algorithm.
For the project, 11,482 citizen science volunteers have processed thousands of images taken by different cameras of the Hubble telescope from 2002 to 2021 and made nearly 1,500 tentative identifications of asteroids in about 1% of the Hubble images provided. Astronomers then used this data to train an automated machine-learning algorithm to search for additional asteroid trails, which completed the analysis of the images.
The researchers note that about a third of the observed asteroid trails can be correlated with known asteroids that have been captured before. But two-thirds of the trails relate to new objects that will require further observation.
Studying these asteroids will help scientists learn more about conditions in the early solar system when planets formed, the project authors say.