Skip to content

Confirmation of Candidate - Candidate : Tomas Vorobjov

Improved Faint Asteroid Discovery Using Deep Learning and Finding Asteroids in an era of Mega-Constellations of Satellites
01 DEC 2021
9.00 AM - 10.30 AM
This research will focus on using deep learning Convolutional Neural Networks to improve candidate detection rates of fainter and faster moving transients in ground based wide-field astronomical survey imagery.
The main goals will be (1) to improve the detection rates of fainter objects which are otherwise either missed or disqualified by classical data processing pipelines (2) to reduce the false positive detections so that more and fainter candidates can be examined by observers, (3) to explore deep learning use to differentiate satellite mega-constellations and asteroids. Efficiency and accuracy of deep learning will be checked using an existing Pan-STARRS Synthetic Solar System Model whereby the Convolutional Neural Networks will be tasked with detecting synthetic transients injected into real survey data.

For more information, please contact the Graduate Research School.