Discovering asteroids in any dataset

Asteroids are typically discovered by taking back-to-back images of the same area of the sky, and looking for objects whose position has shifted. But the vast majority of the world’s astronomical datasets are not taken with this cadence, making them difficult to use for asteroid searches. Algorithms to do so were generally thought to be computationally infeasible.

In 2021, a team lead by UW graduate student and B612 Asteroid Institute Fellow Joachim Moeyens demonstrated an algorithm that can discover asteroid irrespective of the cadence of observations. This novel technique — still under heavy development but already available within ADAM — could significantly increase the efficiency of existing survey programs, and enable researchers to dramatically expand the datasets usable for mapping of the Solar System.

Our team is presently working to generalize THOR to all populations of the Solar System, including NEOs.

• Discover asteroids without the need for tracklets
• 97% efficient at finding asteroids in ZTF, 1.5-2x better than existing codes