Detecting Solar system objects with convolutional neural networks

Abstract

In the preparation for ESA's Euclid mission and the large amount of data it will produce, we train deep convolutional neural networks (CNNs) on Euclid simulations to classify Solar system objects from other astronomical sources. Using transfer learning we are able to achieve a good performance despite our tiny data set with as few as 7512 images. Our best model correctly identifies objects with a top accuracy of 94 per cent and improves to 96 per cent when Euclid's dither information is included. The neural network misses ~50 per cent of the slowest moving asteroids (v < 10 arcsec h-1) but is otherwise able to correctly classify asteroids even down to 26 mag. We show that the same model also performs well at classifying stars, galaxies, and cosmic rays, and could potentially be applied to distinguish all types of objects in the Euclid data and other large optical surveys.