Simulating Realistic Images of the Sky to Help Train Algorithms in Astronomy

Strategic Insights -

Astronomers are drowning in data, with tens of terabytes generated every night by large survey telescopes. To make sense of this massive data deluge, researchers are turning to simulations of realistic images of the sky. By painstakingly simulating light and its interaction with everything it encounters, scientists can create fake but realistic images that help train algorithms to uncover properties of objects in the images. One such simulator, called PhoSim, can simulate the propagation of light particles through the Earth's atmosphere and into the telescope and camera, allowing researchers to understand the origin of systematics in measurements.

Key Takeaways:

  • The Vera Rubin Observatory will take in lots of light to construct maps of the sky, with capabilities to survey the entire southern celestial hemisphere every few nights.
  • The data deluge from large survey telescopes, including the Rubin observatory, is creating a massive problem for astronomers, with tens of terabytes generated every night.
  • PhoSim is a publicly available code that simulates realistic images of the sky by simulating light and its interaction with everything it encounters.
  • The simulator uses a variety of physics to predict what photons do when they encounter the air and the telescope's mirrors and lenses, allowing researchers to understand the origin of systematics in measurements.
  • Researchers have used PhoSim to verify aspects of the Rubin observatory's design and estimate how its images would look.
  • The simulations have connected temperature variations across telescope mirrors to astigmatism in images and studied how high-altitude turbulence in the atmosphere affects image positions and causes blurring patterns.
  • Astronomers can use these new results to correct their measurements and better take advantage of all the data that telescopes collect.

Statistics:

  • The Rubin observatory will survey the entire southern celestial hemisphere every few nights.
  • The observatory will generate tens of terabytes every night.
  • The data deluge from large survey telescopes is creating a massive problem for astronomers.
  • PhoSim simulates the propagation of trillions of photons through the Earth's atmosphere and into the telescope and camera.
  • The simulator uses the Monte Carlo method, which uses random sampling, to represent light as trillions of photons.

Sources:

  • The Conversation -- USA -- By John Peterson, Assoc. Professor of Physics and Astronomy, Purdue University
  • "Astronomy has a major data problem -- simulating realistic images of the sky can help train algorithms" (https://theconversation.com/astronomy-has-a-major-data-problem-simulating-realistic-images-of-the-sky-can-help-train-algorithms-258786)
  • John Peterson/Purdue (simulations of galaxies, photons, and the Rubin observatory)