Seeing is Believing - Astronomy Instrumentation and Machine Learning

People

Supervisory Chair

External Member

Dr Damien Gratadour and Jeffrey Smith

Description

Research state-of-the-art algorithms in statistical machine learning and computer vision and develop a learning system that can accurately estimate astronomical ‘seeing’ conditions.

Large terrestrial telescopes, such as the Extreme Large Telescope (ELT), rely on adaptive optics (AO) systems to compensate for atmospheric distortions. Without these systems, these large telescopes that are approaching 40m in diameter would have the resolving power of a small telescope of some few inches diameter. These AO systems contain a wavefront sensor which has been recently utilised by computer vision techniques at ANU to accurately estimate the wavefront in real time.

Your project is to develop a computer vision system that leverages these new estimation techniques to estimate the Fried parameter – a measure of the astronomical seeing that varies with atmospheric conditions – from wavefront sensor images. This is a highly desirable environmental measurement that is not currently available to astronomers in real time with existing tools. GPU accelerated COMPASS AO loop simulation software is available that can accurately simulate AO for large telescopes to provide experimental data and analyse results.

Goals

Create a computer program that can accurately estimate the Fried parameter from wavefront sensor images directly from AO systems in real time.

Requirements

- Must be comfortable programming and running empirical experiments.

- Familiarity with python and Linux environments.

- Background or interest in Astronomy, Astrophysics and / or Optics and instrumentation is desirable

Gain

Research experience in computer vision and astronomy. The gratitude of astronomers everywhere.

Updated:  10 August 2021/Responsible Officer:  Dean, CECS/Page Contact:  CECS Marketing