Astronomical Image Quality Assessment Based on Deep Learning for Resource-constrained EnvironmentsLi, Juan; Zhang, Xiaoming; Ge, Jiayi; Bai, Chunhai; Feng, Guojie; Mu, Haiyang; Wang, Lei; Liu, Chengzhi; Kang, Zhe; Jiang, Xiaojun
doi: 10.1088/1538-3873/adb790pmid: N/A
This paper presents a highly lightweight model for astronomical image quality assessment, named AQSA-Net, designed to address the challenges of evaluating image quality in scenarios with limited computational resources and rapid decision-making needs. With only 0.15B in computational cost and 0.67M parameters, AQSA-Net significantly reduces memory usage while enhancing inference speed and real-time processing. We construct a data set with eight quality categories based on actual astronomical images. We develop an efficient feature extraction and data processing method that integrates local and global image information, substantially reducing input resolution and training time. We optimize the AQSA-Net architecture and introduce a spatial attention unit, enabling the model to focus on key image areas, enhancing feature extraction while reducing computational overhead. AQSA-Net is compared with several classic deep convolutional neural networks, and experimental results show that AQSA-Net achieves state-of-the-art performance with minimal computational complexity and parameter count. Specifically, AQSA-Net achieves an accuracy of 97.63%, recall of 98.03%, precision of 97.79%, and F1-score of 97.91% on the test set. Additionally, to more accurately assess the quality of usable images, we construct a quantitative image quality factor and a quality grading system, providing quantifiable evaluation criteria for subsequent scientific research. Therefore, our method effectively distinguishes high-quality images from low-quality ones that may impact scientific projects. This provides a reliable automated quality assessment tool for large-scale, complex data sets requiring deep learning inspection. Furthermore, our image quality evaluation could support the assessment of scientific observation data.
Deep Learning-based Detection and Segmentation of Edge-on and Highly Inclined GalaxiesChrobáková, Ž.; Krešňáková, V.; Nagy, R.; Gazdová, J.; Butka, P.
doi: 10.1088/1538-3873/adbcd6pmid: N/A
Edge-on galaxies have many important applications in galactic astrophysics, but they can be difficult to identify in vast amounts of astronomical data. To facilitate the search for them, we developed a deep learning algorithm designed to identify and extract edge-on galaxies from astronomical images. We utilized a sample of edge-on spiral galaxies from the Galaxy Zoo database, retrieving the corresponding images from the Sloan Digital Sky Survey (SDSS). Our data set comprised ∼16,000 galaxies, which we used to train the YOLOv5 algorithm for detection purposes. To isolate galaxies from their backgrounds, we trained the SCSS-Net neural network to generate segmentation masks. As a result, our algorithm detected ∼12,000 edge-on galaxies with a high confidence, for which we compiled a catalog including their parameters obtained from the SDSS database. We described basic properties of our sample, finding that most galaxies have redshifts 0.02 < z < 0.10, have low values of b/a and are mostly red, which is expected from edge-on galaxies and is consistent with our training sample, as well as other literature. The cutouts of the detected galaxies can be used for future studies and the algorithm can be applied to data from future surveys as well.
NCPA Correction Method for AO Systems Based on Phase DiversityLi, Shuqi; Bao, Hua; Bian, Qing; Gao, Guoqing; Zhang, Lulin; Rao, Changhui
doi: 10.1088/1538-3873/adbf49pmid: N/A
Astronomical large aperture high-resolution telescopes adopt adaptive optics (AO) to improve resolution. Due to the physical separation of wavefront detection and imaging optical paths, non-common optical path aberration (NCPA) must be corrected. To correct NCPA, we propose a simple and effective method based on phase diversity (PD) wavefront aberration detection combined with AO closed-loop correction. This technology generates a known defocus aberration through the deformable mirror and collects the focused and defocused spot images of the AO system after closed-loop in a time-division manner. The distorted wavefront aberration is measured by PD algorithm and the corresponding sub-aperture slopes used to modify the wavefront sensor (WFS) reference of the wavefront is calculated by the virtual Hartmann-Shack approach. After several iterations of the AO closed-loop data acquisition, the PD wavefront distortion measurement and the WFS reference modification, the far-field focused spot will gradually converge to the diffraction limit Airy spot of the imaging system. The NCPAs of 705.6 nm and 656.3 nm bands in the solar adaptive optics system of the 1.8 m Chinese Large Solar Telescope are measured and corrected in this article, and the strehl ration values of focused spot images increase from 0.689 and 0.373 to 0.839 and 0.770, corresponding to the rms values of compensated NCPA wavefront aberrations being 50.05 nm and 72.91 nm respectively. The experiments verify the effectiveness of the proposed method, and the results show that this method has extremely high NCPA correction accuracy, which has important application value in fields such as solar AO system and extreme AO system.
ELT-METIS Imaging Simulations for Disks and Envelopes Associated with FU Ori-type ObjectsTakami, Michihiro; Otten, Gilles; Absil, Olivier; Delacroix, Christian; Karr, Jennifer L.; Wang, Shiang-Yu
doi: 10.1088/1538-3873/adbbc4pmid: N/A
We investigate the detectability of extended mid-infrared (MIR) emission associated with FU-Ori type objects (FUors) using the Mid-infrared ELT Imager and Spectrograph (METIS) coronagraphs on the 39 m Extremely Large Telescope. The imaging simulations were made for three representative filters (λ = 3.8, 4.8, and 11.3 μm) of the METIS instrument. We demonstrate that the detectability of the extended MIR emission using these coronagraphs is highly dependent on the uncertain nature of the central FUor and its circumstellar environment in various contexts. These contexts are: (A) whether the central radiation source is either a flat self-luminous accretion disk or a star at near-infrared (NIR) wavelengths, (B) the size of the accretion disk for the bright central MIR emission at milliarcsecond scales, (C) whether the extended emission is due to either an optically thick disk or an optically thin envelope, and (D) dust grain models. Observations at λ = 3.8 μm will allow us to detect the extended emission in many cases, while the number of cases with detection may significantly decrease toward longer wavelengths due to the fainter nature of the extended emission and high thermal background noise. In some cases, the presence of a binary companion can significantly hamper detections of the extended MIR emission. NIR and MIR imaging observations at existing 8 m class telescopes, prior to the METIS observations, will be useful for (1) reducing the many model uncertainties and (2) searching for binary companions associated with FUors, therefore determining the best observing strategy using METIS.
CARRSSPipeline: Flux Calibration and Nonlinear Reprojection for SALT-RSS Multi-Object Spectroscopy over 3500–9500 ÅKharchilava, George V.; Gawiser, Eric; Hilton, Matt; Turner, Elisabeth; Firestone, Nicole M.; Lee, Kyoung-Soo
doi: 10.1088/1538-3873/adb454pmid: N/A
The Robert Stobie Spectrograph (RSS) on the Southern African Large Telescope (SALT) offers multi-object spectroscopy over an 8′ field-of-view at resolutions up to R ∼ 3000. Reduction is typically conducted using RSSMOSPipeline, which performs basic data calibrations, sky subtraction, and wavelength calibration. However, flux calibration of SALT-RSS using spectrophotometric standard star observations is difficult due to variable primary mirror illumination. We describe a novel approach where stars with Sloan Digital Sky Survey spectra are included as alignment stars on RSS slitmasks and then used to perform a rough flux calibration of the resulting data. RSS offers multiple settings that can be pieced together to cover the entire optical range, utilizing grating angle dithers to fill chip gaps. We introduce a nonlinear reprojection routine that defines an exponential wavelength array spanning 3500–9500 Å with gradually decreasing resolution and then reprojects several individual settings into a single 2D spectrum for each object. Our flux calibration and nonlinear reprojection routines are released as part of the Calibration And Reprojection for RSS Pipeline (CARRSSPipeline), that enables the extraction of full-optical-coverage, flux-calibrated, medium-resolution one-dimensional spectra.
Mixed Origins: Strong Natal Kicks for Some Black Holes and None for OthersNagarajan, Pranav; El-Badry, Kareem
doi: 10.1088/1538-3873/adb6d6pmid: N/A
Using stellar kinematic data from Gaia DR3, we revisit constraints on black hole (BH) natal kicks from observed accreting and detached BH binaries. We compare the space velocities and Galactic orbits of a sample of 12 BHs in the Galactic disk with well-constrained distances to their local stellar populations, for which we obtain proper motions and radial velocities from Gaia DR3. Compared to most previous studies, we infer lower minimum kick velocities, because our modeling accounts for the fact that most BH binaries are old and have likely been kinematically heated by processes other than kicks. Nevertheless, we find that half of the BHs have at least weak evidence for a kick, being kinematically hotter than at least 68% of their local stellar populations. At least 4 BHs are kinematically hotter than 90% of their local stellar populations, suggesting they were born with kicks of ≳100 km s−1. On the other hand, 6 BHs have kinematics typical of their local populations, disfavoring kicks of ≳50 km s−1. For two BHs, V404 Cyg and VFTS 243, there is strong independent evidence for a very weak kick ≲10 km s−1. Our analysis implies that while some BHs must form with very weak kicks, it would be wrong to conclude that most BHs do, particularly given that selection biases favor weak kicks. Although the uncertainties on most individual BHs’ kicks are still too large to assess whether the kick distribution is bimodal, the data are consistent with a scenario where some BHs form by direct collapse and receive weak kicks, and others form in supernovae and receive strong kicks.
The Effect of Vera C. Rubin Observatory Cadence Selections on Kilonova DetectabilityAndrade, Cristina; Alserkal, Raiyah; Salazar Manzano, Luis; Martin, Emma; Andreoni, Igor; Coughlin, Michael W.; Guessoum, Nidhal; Rivera Sandoval, Liliana
doi: 10.1088/1538-3873/adbfbcpmid: N/A
The discovery of the optical/infra-red counterpart (AT2017gfo) to the binary neutron star gravitational-wave detection (GW170817), which was followed by a short gamma-ray burst (GRB 170817), marked a groundbreaking moment in multi-messenger astronomy. To date, it remains the only confirmed joint detection of its kind. However, many experiments are actively searching for similar fast-fading electromagnetic counterparts, known as kilonovae. Fortunately, the Vera C. Rubin Observatory’s Legacy Survey of Space and Time (LSST) provides excellent prospects for identifying kilonova candidates either from, or independent of, gravitational-wave and GRB triggers. Cadence choices for LSST surveys are especially important for maximising the likelihood of kilonovae detections. In this work, we explore the possibility of optimizing Rubin Observatory’s ability to detect kilonovae by implementing a fast transient metric shown to be successful with an existing wide field survey, e.g., the Zwicky Transient Facility. We study existing LSST cadences, how detection rates are affected by filter selections, the return timescales for visits of the same area in the sky, and other relevant factors. Through our analysis, we have found that employing baseline cadences and utilizing triplet families like presto_gap produced the highest likelihood of kilonova detection.