Taxonomy of Asteroids From the Legacy Survey of Space and Time Using Neural Networks

Abstract

Read online

The Legacy Survey of Space and Time (LSST) is one of the ongoing or future surveys, together with the Gaia and Euclid missions, which will produce a wealth of spectrophotometric observations of asteroids. This article shows how deep learning techniques with neural networks can be used to classify the upcoming observations, particularly from LSST, into the Bus-DeMeo taxonomic system. We report here a success ratio in classification up to 90.1% with a reduced set of Bus-DeMeo types for simulated observations using the LSST photometric filters. The scope of this work is to introduce tools to link future observations into existing Bus-DeMeo taxonomy.

Keywords

• legacy survey of space and time
• asteroids
• spectra
• taxonomy
• neural network