Repozitorij Univerze v Novi Gorici

Iskanje po repozitoriju
A+ | A- | Pomoč | SLO | ENG

Iskalni niz: išči po
išči po
išči po
išči po
* po starem in bolonjskem študiju


1 - 4 / 4
Na začetekNa prejšnjo stran1Na naslednjo stranNa konec
AutoSourceID-Classifier : star-galaxy classification using a convolutional neural network with spatial information
F. Stoppa, Saptashwa Bhattacharyya, R. Ruiz de Austri, P. Vreeswijk, S. Caron, Gabrijela Zaharijas, S. Bloemen, G. Principe, D. Malyshev, Veronika Vodeb, 2023, izvirni znanstveni članek

Opis: Aims: Traditional star-galaxy classification techniques often rely on feature estimation from catalogs, a process susceptible to introducing inaccuracies, thereby potentially jeopardizing the classification’s reliability. Certain galaxies, especially those not manifesting as extended sources, can be misclassified when their shape parameters and flux solely drive the inference. We aim to create a robust and accurate classification network for identifying stars and galaxies directly from astronomical images. Methods: The AutoSourceID-Classifier (ASID-C) algorithm developed for this work uses 32x32 pixel single filter band source cutouts generated by the previously developed AutoSourceID-Light (ASID-L) code. By leveraging convolutional neural networks (CNN) and additional information about the source position within the full-field image, ASID-C aims to accurately classify all stars and galaxies within a survey. Subsequently, we employed a modified Platt scaling calibration for the output of the CNN, ensuring that the derived probabilities were effectively calibrated, delivering precise and reliable results. Results: We show that ASID-C, trained on MeerLICHT telescope images and using the Dark Energy Camera Legacy Survey (DECaLS) morphological classification, is a robust classifier and outperforms similar codes such as SourceExtractor. To facilitate a rigorous comparison, we also trained an eXtreme Gradient Boosting (XGBoost) model on tabular features extracted by SourceExtractor. While this XGBoost model approaches ASID-C in performance metrics, it does not offer the computational efficiency and reduced error propagation inherent in ASID-C’s direct image-based classification approach. ASID-C excels in low signal-to-noise ratio and crowded scenarios, potentially aiding in transient host identification and advancing deep-sky astronomy.
Ključne besede: astronomical databases, data analysis, statistics, image processing
Objavljeno v RUNG: 12.12.2023; Ogledov: 771; Prenosov: 4
.pdf Celotno besedilo (10,31 MB)
Gradivo ima več datotek! Več...

AutoSourceID-FeatureExtractor : optical image analysis using a two-step mean variance estimation network for feature estimation and uncertainty characterisation
F. Stoppa, R. Ruiz de Austri, P. Vreeswijk, Saptashwa Bhattacharyya, S. Caron, S. Bloemen, Gabrijela Zaharijas, G. Principe, Veronika Vodeb, P. J. Groot, E. Cator, G. Nelemans, 2023, izvirni znanstveni članek

Opis: Aims: In astronomy, machine learning has been successful in various tasks such as source localisation, classification, anomaly detection, and segmentation. However, feature regression remains an area with room for improvement. We aim to design a network that can accurately estimate sources' features and their uncertainties from single-band image cutouts, given the approximated locations of the sources provided by the previously developed code AutoSourceID-Light (ASID-L) or other external catalogues. This work serves as a proof of concept, showing the potential of machine learning in estimating astronomical features when trained on meticulously crafted synthetic images and subsequently applied to real astronomical data. Methods: The algorithm presented here, AutoSourceID-FeatureExtractor (ASID-FE), uses single-band cutouts of 32x32 pixels around the localised sources to estimate flux, sub-pixel centre coordinates, and their uncertainties. ASID-FE employs a two-step mean variance estimation (TS-MVE) approach to first estimate the features and then their uncertainties without the need for additional information, for example the point spread function (PSF). For this proof of concept, we generated a synthetic dataset comprising only point sources directly derived from real images, ensuring a controlled yet authentic testing environment. Results: We show that ASID-FE, trained on synthetic images derived from the MeerLICHT telescope, can predict more accurate features with respect to similar codes such as SourceExtractor and that the two-step method can estimate well-calibrated uncertainties that are better behaved compared to similar methods that use deep ensembles of simple MVE networks. Finally, we evaluate the model on real images from the MeerLICHT telescope and the Zwicky Transient Facility (ZTF) to test its transfer learning abilities.
Ključne besede: data analysis, image processing, astronomical databases
Objavljeno v RUNG: 08.11.2023; Ogledov: 744; Prenosov: 7
URL Povezava na datoteko
Gradivo ima več datotek! Več...

AutoSourceID-Light : Fast optical source localization via U-Net and Laplacian of Gaussian
F. Stoppa, P. Vreeswijk, S. Bloemen, Saptashwa Bhattacharyya, S Caron, G. Jóhannesson, R. Ruiz de Austri, C. Van den Oetelaar, Gabrijela Zaharijas, P.J. Groot, E. Cator, G. Nelemans, 2022, izvirni znanstveni članek

Opis: Aims: With the ever-increasing survey speed of optical wide-field telescopes and the importance of discovering transients when they are still young, rapid and reliable source localization is paramount. We present AutoSourceID-Light (ASID-L), an innovative framework that uses computer vision techniques that can naturally deal with large amounts of data and rapidly localize sources in optical images. Methods: We show that the ASID-L algorithm based on U-shaped networks and enhanced with a Laplacian of Gaussian filter provides outstanding performance in the localization of sources. A U-Net network discerns the sources in the images from many different artifacts and passes the result to a Laplacian of Gaussian filter that then estimates the exact location. Results: Using ASID-L on the optical images of the MeerLICHT telescope demonstrates the great speed and localization power of the method. We compare the results with SExtractor and show that our method outperforms this more widely used method rapidly detects more sources not only in low and mid-density fields, but particularly in areas with more than 150 sources per square arcminute. The training set and code used in this paper are publicly available.
Ključne besede: astronomical databases, data analysis, image processing
Objavljeno v RUNG: 23.01.2023; Ogledov: 1475; Prenosov: 0
Gradivo ima več datotek! Več...

Tidal Disruption Events seen through the eyes of Vera C. Rubin Observatory
Katja Bučar Bricman, 2021, doktorska disertacija

Opis: Tidal Disruption Events (TDEs) are rare transients, which are considered to be promising tools in probing supermassive black holes (SMBHs) and their environments in quiescent galaxies, accretion physics, and jet formation mechanisms. The majority of $\approx$ 60 detected TDEs has been discovered with large field of view time-domain surveys in the last two decades. Currently, about 10 TDEs are discovered per year, and we expect this number will increase largely once the Legacy Survey of Space and Time (LSST) at Vera C. Rubin Observatory begins its observations. In this work we demonstrate and explore the capabilities of the LSST to study TDEs. To begin with, we simulate LSST observations of TDEs over $10$ years of survey duration by including realistic SED models from MOSFiT into the simulation framework of the LSST. SEDs are then converted into observed fluxes and light curves are simulated with the LSST observing strategy minion_1016. Simulated observations are used to estimate the number of TDEs the LSST is expected to observe and to assess the possibility of probing the SMBH mass distribution in the Universe with the observed TDE sample. We find that the LSST has a potential of observing ~1000 TDEs per year, the exact number depending on the SMBH mass distribution and the adopted observing strategy. In spite of this large number, we find that probing the SMBH mass distribution with LSST observed TDEs will not be straightforward, especially at the low-mass end. This is largely attributed to the fact that TDEs caused by low-mass black holes ($\le 10^6 M_\odot$) are less luminous and shorter than TDEs by heavier SMBHs ($> 10^6 M_\odot$), and the probability of observationally missing them with LSST is higher. Second, we built a MAF TDE metric for photometric identification of TDEs based on LSST data. We use the metric to evaluate the performance of different proposed survey strategies in identifying TDEs with pre-defined identification requirements. Since TDEs are blue in color for months after peak light, which separates them well from SNe and AGN, we include u-band observations as one of the criteria for a positive identification. We find that the number of identified TDEs strongly depends of the observing strategy and the number of u-band visits to a given field in the sky. Observing strategies with a larger number of u-band observations perform significantly better. For these strategies up to 10% of LSST observed TDEs satisfy the identification requirements.
Ključne besede: Ground-based ultraviolet, optical and infrared telescopes Astronomical catalogs, atlases, sky surveys, databases, retrieval systems, archives, Black holes, Galactic nuclei (including black holes), circumnuclear matter, and bulges, Infall, accretion, and accretion disks
Objavljeno v RUNG: 03.01.2022; Ogledov: 2994; Prenosov: 66
.pdf Celotno besedilo (124,61 MB)

Iskanje izvedeno v 0.03 sek.
Na vrh