Learning to denoise astronomical images with U-nets

Kód výsledku v IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F00216224%3A14310%2F21%3A00123743" target="_blank" >RIV/00216224:14310/21:00123743 - isvavai.cz</a>

Výsledek na webu

<a href="https://doi.org/10.1093/mnras/staa3567" target="_blank" >https://doi.org/10.1093/mnras/staa3567</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1093/mnras/staa3567" target="_blank" >10.1093/mnras/staa3567</a>

Jazyk výsledku

angličtina

Název v původním jazyce

Learning to denoise astronomical images with U-nets

Popis výsledku v původním jazyce

Y Astronomical images are essential for exploring and understanding the Universe. Optical telescopes capable of deep observations, such as the Hubble Space Telescope (HST), are heavily oversubscribed in the Astronomical Community. Images also often contain additive noise, which makes denoising a mandatory step in post-processing the data before further data analysis. In order to maximize the efficiency and information gain in the post-processing of astronomical imaging, we turn to machine learning. We propose ASTRO U-NET, a convolutional neural network for image denoising and enhancement. For a proof-of-concept, we use HST images from Wide Field Camera 3 instrument UV/visible channel with F555W and F606W filters. Our network is able to produce images with noise characteristics as if they are obtained with twice the exposure time, and with minimum bias or information loss. From these images, we are able to recover 95.9 per cent of stars with an average flux error of 2.26 per cent. Furthermore, the images have, on average, 1.63 times higher signal-to-noise ratio than the input noisy images, equivalent to the stacking of at least three input images, which means a significant reduction in the telescope time needed for future astronomical imaging campaigns.

Název v anglickém jazyce

Learning to denoise astronomical images with U-nets

Popis výsledku anglicky

Y Astronomical images are essential for exploring and understanding the Universe. Optical telescopes capable of deep observations, such as the Hubble Space Telescope (HST), are heavily oversubscribed in the Astronomical Community. Images also often contain additive noise, which makes denoising a mandatory step in post-processing the data before further data analysis. In order to maximize the efficiency and information gain in the post-processing of astronomical imaging, we turn to machine learning. We propose ASTRO U-NET, a convolutional neural network for image denoising and enhancement. For a proof-of-concept, we use HST images from Wide Field Camera 3 instrument UV/visible channel with F555W and F606W filters. Our network is able to produce images with noise characteristics as if they are obtained with twice the exposure time, and with minimum bias or information loss. From these images, we are able to recover 95.9 per cent of stars with an average flux error of 2.26 per cent. Furthermore, the images have, on average, 1.63 times higher signal-to-noise ratio than the input noisy images, equivalent to the stacking of at least three input images, which means a significant reduction in the telescope time needed for future astronomical imaging campaigns.

Druh

J_imp - Článek v periodiku v databázi Web of Science

CEP obor

—

OECD FORD obor

10308 - Astronomy (including astrophysics,space science)

Projekt

—

Návaznosti

S - Specificky vyzkum na vysokych skolach<br>I - Institucionalni podpora na dlouhodoby koncepcni rozvoj vyzkumne organizace

Rok uplatnění

2021

Kód důvěrnosti údajů

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Název periodika

Monthly Notices of the Royal Astronomical Society

ISSN

0035-8711

e-ISSN

1365-2966

Svazek periodika

503

Číslo periodika v rámci svazku

3

Stát vydavatele periodika

GB - Spojené království Velké Británie a Severního Irska

Počet stran výsledku

12

Strana od-do

3204-3215

Kód UT WoS článku

000649000600006

EID výsledku v databázi Scopus

2-s2.0-85110340238