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 $95.9{{\ \rm per\ cent}}$ of stars with an average flux error of $2.26{{\ \rm per\ cent}}$. 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.
19 December 2019 in Monthly Notices of the Royal Astronomical Society
Scaling relations trace the formation and evolution of galaxy clusters. We exploited multi-wavelength surveys – the XXL survey at XMM-Newton in the X-ray band, and the Hyper Suprime-Cam Subaru Strategic Program for optical weak lensing – to study an X-ray selected, complete sample of clusters and groups. The scalings of gas mass, temperature, and soft-band X-ray luminosity with the weak lensing mass show imprints of radiative cooling and active galactic nucleus feedback in groups. From the multi-variate analysis, we found some evidence for steeper than self-similar slopes for gas mass ($\beta _{m_\text{g}|m}=1.73 \pm 0.80$) and luminosity (βl|m = 1.91 ± 0.94) and a nearly self-similar slope for the temperature (βt|m = 0.78 ± 0.43). Intrinsic scatters of X-ray properties appear to be positively correlated at a fixed mass (median correlation factor $\rho _{X_1X_2|m}\sim 0.34$) due to dynamical state and merger history of the haloes. Positive correlations with the weak lensing mass (median correlation factor $\rho _{m_\text{wl}X|m}\sim 0.35$) can be connected to triaxiality and orientation. Comparison of weak lensing and hydrostatic masses suggests a small role played by non-thermal pressure support ($9\pm 17{{\ \rm per\ cent}}$).
14 March 2019 in Monthly Notices of the Royal Astronomical Society
In the preparation for ESA’s Euclid mission and the large amount of data it will produce, we train deep convolutional neural networks on Euclid simulations to classify solar system objects from other astronomical sources. Using transfer learning we are able to achieve a good performance despite our tiny dataset with as few as 7512 images. Our best model correctly identifies objects with a top accuracy of 94% and improves to 96% when Euclid’s dither information is included. The neural network misses ∼50% of the slowest moving asteroids (v < 10 arcsec h−1) but is otherwise able to correctly classify asteroids even down to 26 mag. We show that the same model also performs well at classifying stars, galaxies and cosmic rays, and could potentially be applied to distinguish all types of objects in the Euclid data and other large optical surveys.
8 September 2022 in Monthly Notices of the Royal Astronomical Society
The field of artificial intelligence based image enhancement has been rapidly evolving over the last few years and is able to produce impressive results on non-astronomical images. In this work we present the first application of Machine Learning based super-resolution (SR) and de-noising (DN) to enhance X-ray images from the European Space Agency’s XMM-Newton telescope. Using XMM-Newton images in band [0.5,2] keV from the European Photon Imaging Camera pn detector (EPIC-pn), we develop XMM-SuperRes and XMM-DeNoise — deep learning-based models that can generate enhanced SR and DN images from real observations. The models are trained on realistic XMM-Newton simulations such that XMM-SuperRes will output images with two times smaller point-spread function and with improved noise characteristics. The XMM-DeNoise model is trained to produce images with 2.5× the input exposure time from 20 to 50 ks. When tested on real images, DN improves the image quality by |$8.2{{\%}}$|, as quantified by the global peak-signal-to-noise ratio. These enhanced images allow identification of features that are otherwise hard or impossible to perceive in the original or in filtered/smoothed images with traditional methods. We demonstrate the feasibility of using our deep learning models to enhance XMM-Newton X-ray images to increase their scientific value in a way that could benefit the legacy of the XMM-Newton archive.
1 August 2022 in Monthly Notices of the Royal Astronomical Society
We present a new X-Ray Accretion Disc-wind Emulator (xrade) based on the 2.5D Monte Carlo radiative transfer code that provides a physically motivated, self-consistent treatment of both absorption and emission from a disc wind by computing the local ionization state and velocity field within the flow. xrade is then implemented through a process that combines X-ray tracing with supervised machine learning. We develop a novel emulation method consisting in training, validating, and testing the simulated disc-wind spectra into a purposely built artificial neural network. The trained emulator can generate a single synthetic spectrum for a particular parameter set in a fraction of a second, in contrast to the few hours required by a standard Monte Carlo radiative transfer pipeline. The emulator does not suffer from interpolation issues with multidimensional spaces that are typically faced by traditional X-ray fitting packages such as xspec. xrade will be suitable to a wide number of sources across the black hole mass, ionizing luminosity, and accretion rate scales. As an example, we demonstrate the applicability of xrade to the physical interpretation of the X-ray spectra of the bright quasar PDS 456, which hosts the best-established accretion disc wind observed to date. We anticipate that our emulation method will be an indispensable tool for the development of high-resolution theoretical models, with the necessary flexibility to be optimized for the next generation microcalorimeters onboard future missions, like X-Ray Imaging and Spectroscopy Mission (XRISM)/Resolve and Athena/X-ray Integral Field Unit (X-IFU). This tool can also be implemented across a wide variety of X-ray spectral models and beyond.
22 March 2017 in Monthly Notices of the Royal Astronomical Society
Mass is a fundamental property of galaxy groups and clusters. In principle, weak gravitational lensing will enable an approximately unbiased measurement of mass, but parametric methods for extracting cluster masses from data require the additional knowledge of halo concentration. Measurements of both mass and concentration are limited by the degeneracy between the two parameters, particularly in low-mass, high-redshift systems where the signal to noise is low. In this paper, we develop a hierarchical model of mass and concentration for mass inference, we test our method on toy data and then apply it to a sample of galaxy groups and poor clusters down to masses of ∼ 1013 M⊙. Our fit and model gives a relationship among masses, concentrations and redshift that allow prediction of these parameters from incomplete and noisy future measurements. Additionally, the underlying population can be used to infer an observationally based concentration–mass relation. Our method is equivalent to a quasi-stacking approach with the degree of stacking set by the data. We also demonstrate that mass and concentration derived from pure stacking can be offset from the population mean with differing values depending on the method of stacking.