Cosmic Ray Hits Research Articles

A Comparison of Neural Networks and Center of Gravity in Muon Hit Position Estimation.

The performance of cosmic-ray tomography systems is largely determined by their tracking accuracy. With conventional scintillation detector technology, good precision can be achieved with a small pitch between the elements of the detector array. Improving the resolution implies increasing the number of read-out channels, which in turn increases the complexity and cost of the tracking detectors. As an alternative to that, a scintillation plate detector coupled with multiple silicon photomultipliers could be used as a technically simple solution. In this paper, we present a comparison between two deep-learning-based methods and a conventional Center of Gravity (CoG) algorithm, used to calculate cosmic-ray muon hit positions on the plate detector using the signals from the photomultipliers. In this study, we generated a dataset of muon hits on a detector plate using the Monte Carlo simulation toolkit GEANT4. We demonstrate that two deep-learning-based methods outperform the conventional CoG algorithm by a significant margin. Our proposed algorithm, Fully Connected Network, produces a 0.72 mm average error measured in Euclidean distance between the actual and predicted hit coordinates, showing great improvement in comparison with CoG, which yields 1.41 mm on the same dataset. Additionally, we investigated the effects of different sensor configurations on performance.

Cosmic Ray rejection with attention augmented deep learning

Cosmic Ray (CR) hits are the major contaminants in astronomical imaging and spectroscopic observations involving solid-state detectors. Correctly identifying and masking them is a crucial part of the image processing pipeline, since it may otherwise lead to spurious detections. For this purpose, we have developed and tested a novel Deep Learning based framework for the automatic detection of CR hits from astronomical imaging data from two different imagers: Dark Energy Camera (DECam) and Las Cumbres Observatory Global Telescope (LCOGT). We considered two baseline models namely deepCR and Cosmic-CoNN, which are the current state-of-the-art learning based algorithms that were trained using Hubble Space Telescope (HST) ACS/WFC and LCOGT Network images respectively. We have experimented with the idea of augmenting the baseline models using Attention Gates (AGs) to improve the CR detection performance. We have trained our models on DECam data and demonstrate a consistent marginal improvement by adding AGs in True Positive Rate (TPR) at 0.01% False Positive Rate (FPR) and Precision at 95% TPR over the aforementioned baseline models for the DECam dataset. We demonstrate that the proposed AG augmented models provide significant gain in TPR at 0.01% FPR when tested on previously unseen LCO test data having images from three distinct telescope classes. Furthermore, we demonstrate that the proposed baseline models with and without attention augmentation outperform state-of-the-art models such as Astro-SCRAPPY, Maximask (that is trained natively on DECam data) and pre-trained ground-based Cosmic-CoNN. This study demonstrates that the AG module augmentation enables us to get a better deepCR and Cosmic-CoNN models and to improve their generalization capability on unseen data.

Total power horn-coupled 150 GHz LEKID array for space applications

We have developed two arrays of lumped element kinetic inductance detectors working in the D-band, and optimised for the low radiative background conditions of a satellite mission aiming at precision measurements of the Cosmic Microwave Background (CMB) radiation. The first detector array is sensitive to the total power of the incoming radiation to which is coupled via single-mode waveguides and corrugated feed-horns, while the second is sensitive to the polarisation of the radiation thanks to orthomode transducers.Here, we focus on the total power detector array, which is suitable, for instance, for precision measurements of unpolarised spectral distortions of the CMB, where detecting both polarisations provides a sensitivity advantage. We describe the optimisation of the array design, fabrication and packaging, the dark and optical characterisation, and the performance of the black-body calibrator used for the optical tests. We show that almost all the detectors of the array are photon-noise limited under the radiative background of a 3.6 K black-body. This result, combined with the weak sensitivity to cosmic ray hits demonstrated with the OLIMPO flight, validates the idea of using lumped elements kinetic inductance detectors for precision, space-based CMB missions.

A machine learning approach to the detection of ghosting and scattered light artifacts in dark energy survey images

Astronomical images are often plagued by unwanted artifacts that arise from a number of sources including imperfect optics, faulty image sensors, cosmic ray hits, and even airplanes and artificial satellites. Spurious reflections (known as "ghosts") and the scattering of light off the surfaces of a camera and/or telescope are particularly difficult to avoid. Detecting ghosts and scattered light efficiently in large cosmological surveys that will acquire petabytes of data can be a daunting task. In this paper, we use data from the Dark Energy Survey to develop, train, and validate a machine learning model to detect ghosts and scattered light using convolutional neural networks. The model architecture and training procedure is discussed in detail, and the performance on the training and validation set is presented. Testing is performed on data and results are compared with those from a ray-tracing algorithm. As a proof of principle, we have shown that our method is promising for the Rubin Observatory and beyond.

Application of Neural Networks to Classification of Data of the TUS Orbital Telescope

We employ neural networks for classification of data of the TUS fluorescence telescope, the world’s first orbital detector of ultra-high energy cosmic rays. We focus on two particular types of signals in the TUS data: track-like flashes produced by cosmic ray hits of the photodetector and flashes that originated from distant lightnings. We demonstrate that even simple neural networks combined with certain conventional methods of data analysis can be highly effective in tasks of classification of data of fluorescence telescopes.

Recognition of Cosmic Ray Images Obtained from CMOS Sensors Used in Mobile Phones by Approximation of Uncertain Class Assignment with Deep Convolutional Neural Network.

In this paper, we describe the convolutional neural network (CNN)-based approach to the problems of categorization and artefact reduction of cosmic ray images obtained from CMOS sensors used in mobile phones. As artefacts, we understand all images that cannot be attributed to particles’ passage through sensor but rather result from the deficiencies of the registration procedure. The proposed deep neural network is composed of a pretrained CNN and neural-network-based approximator, which models the uncertainty of image class assignment. The network was trained using a transfer learning approach with a mean squared error loss function. We evaluated our approach on a data set containing 2350 images labelled by five judges. The most accurate results were obtained using the VGG16 CNN architecture; the recognition rate (RR) was 85.79% ± 2.24% with a mean squared error (MSE) of 0.03 ± 0.00. After applying the proposed threshold scheme to eliminate less probable class assignments, we obtained a RR of 96.95% ± 1.38% for a threshold of 0.9, which left about 62.60% ± 2.88% of the overall data. Importantly, the research and results presented in this paper are part of the pioneering field of the application of citizen science in the recognition of cosmic rays and, to the best of our knowledge, this analysis is performed on the largest freely available cosmic ray hit dataset.

MAXIMASK and MAXITRACK: Two new tools for identifying contaminants in astronomical images using convolutional neural networks

In this work, we propose two convolutional neural network classifiers for detecting contaminants in astronomical images. Once trained, our classifiers are able to identify various contaminants, such as cosmic rays, hot and bad pixels, persistence effects, satellite or plane trails, residual fringe patterns, nebulous features, saturated pixels, diffraction spikes, and tracking errors in images. They encompass a broad range of ambient conditions, such as seeing, image sampling, detector type, optics, and stellar density. The first classifier, MAXIMASK, performs semantic segmentation and generates bad pixel maps for each contaminant, based on the probability that each pixel belongs to a given contaminant class. The second classifier, MAXITRACK, classifies entire images and mosaics, by computing the probability for the focal plane to be affected by tracking errors. We gathered training and testing data from real data originating from various modern charged-coupled devices and near-infrared cameras, that are augmented with image simulations. We quantified the performance of both classifiers and show that MAXIMASK achieves state-of-the-art performance for the identification of cosmic ray hits. Thanks to a built-in Bayesian update mechanism, both classifiers can be tuned to meet specific science goals in various observational contexts.

High-cadence optical transient searches using drift scan imaging I: Proof of concept with a pre-prototype system

AbstractAn imaging technique with sensitivity to short duration optical transients is described. The technique is based on the use of wide-field cameras operating in a drift scanning mode, whereby persistent objects produce trails on the sensor and short duration transients occupy localised groups of pixels. A benefit of the technique is that sensitivity to short duration signals is not accompanied by massive data rates, because the exposure time is much greater than the transient duration. The technique is demonstrated using a pre-prototype system composed of readily available and inexpensive commercial components, coupled with common coding environments, commercially available software, and free web-based services. The performance of the technique and the pre-prototype system is explored, including aspects of photometric and astrometric calibration, detection sensitivity, characterisation of candidate transients, and the differentiation of astronomical signals from non-astronomical signals (primarily glints from satellites in Earth orbit and cosmic ray hits on sensor pixels). Test observations were made using the pre-prototype system, achieving sensitivity to transients with 21-ms duration, resulting in the detection of five candidate transients. An investigation of these candidates concludes they are most likely due to cosmic ray hits on the sensor and/or satellites. The sensitivity obtained with the pre-prototype system is such that, under some models for the optical emission from fast radio bursts (FRBs), the detection of a typical FRB, such as FRB181228, to a distance of approximately 100 Mpc is plausible. Several options for improving the system/technique in the future are described.

The Breakthrough Listen Search for Intelligent Life: Searching Boyajian's Star for Laser Line Emission

Boyajian's Star (KIC 8462852) has received significant attention, due to its unusual periodic brightness fluctuations detected by the Kepler spacecraft and subsequent ground-based observations. Possible explanations for the dips in the photometric measurements include interstellar or circumstellar dust, and it has been speculated that an artificial megastructure could be responsible. We analyze 177 high-resolution spectra of Boyajian’s Star in an effort to detect potential laser signals from extraterrestrial civilizations. The spectra were obtained by the Lick Observatory’s Automated Planet Finder (APF) telescope as part of the Breakthrough Listen Project, and cover the wavelength range of visible light from 374 to 970 nm. We calculate that the APF would be capable of detecting lasers of power greater than approximately 24 MW at the distance of Boyajian’s Star, d = 1470 ly. The top candidates from the analysis can all be explained as either cosmic-ray hits, stellar emission lines or atmospheric airglow emission lines.

Hot star extension to theHubbleSpace Telescope stellar spectral library

Context.Libraries of stellar spectra find many uses in astrophysics, from photometric calibration to stellar population synthesis.Aims.We present low resolution spectra of 40 stars from 0.2μm (ultraviolet) to 1.0μm (near infrared) with excellent fluxing. The stars include normal O-type stars, helium-burning stars, and post-asymptotic giant branch (PAGB) stars.Methods.Spectra were obtained with the Space Telescope Imaging Spectrograph (STIS) installed in theHubbleSpace Telescope (HST) using three low resolution gratings, G230LB, G430L, and G750L. Cosmic ray hits and fringing in the red were corrected. A correction for scattered light was applied, significant only for our coolest stars. Cross-correlation was used to bring the spectra to a common, final, zero velocity wavelength scale. Finally, synthetic stellar spectra were used to estimate line of sight dust extinction to each star, and a five-parameter dust extinction model was fit, or a one-parameter fit in the case of low extinction.Results.These spectra dovetail with the similar Next Generation Stellar Library (NGSL) spectra, extending the NGSL’s coverage of stellar parameters, and extending to helium burning stars and stars that do not fuse.Conclusions.The fitted dust extinction model showed considerable variation from star to star, indicating variations in dust properties for different lines of sight. Interstellar absorption lines are present in most stars, notably MgII.

A study of cosmic ray flux based on the noise in raw CCD data from solar images

AbstractRaw solar images from CCDs are often contaminated with single‐pixel noise which is thought to be made by cosmic ray hits. The cosmic ray‐affected pixels are usually outstanding when compared with the perimeter zone. In this work, we use a method based on the median filtering algorithm to identify and count the cosmic ray traces from SOHO/EIT solar images to estimate the cosmic ray (CR) flux. With such cosmic ray flux, we study the transient variations associated with the violent solar activities, such as the solar proton events (SPEs), which show good similarity with the observations of GOES 11 P6 channel with an energy interval 80–165 MeV. Further, using SPE list observed by SOHO/ERNE proton channels with more narrow energy intervals, it is found that CRs in the energy range 118–140 MeV affect the SOHO/EIT images the most. In addition, by using a robust automatic despiking method, we get the background of the cosmic ray flux from solar images, which is considered to be the galactic cosmic ray (GCR) flux. The GCR flux from solar images shows an 11 year period due to the solar modulation, similar to the SOHO/ERNE GCR flux and Newark neutron monitor count rates. Furthermore, GCRs from solar images have a 27 day period and show good anticorrelation with the changes of solar wind velocity.

Correcting Systematic Polarization Effects in Keck LRISp Spectropolarimetry to < 0.05%

Spectropolarimetric measurements at moderate spectral resolutions are effective tracers of stellar magnetic fields and circumstellar environments when signal-to-noise ratios (S/Ns) above 2000 can be achieved. The LRISp spectropolarimeter is capable of achieving these S/Ns on faint targets with the 10 m aperture of the Keck telescope, provided several instrumental artifacts can be suppressed. We describe here several methods to overcome instrumental error sources that are required to achieve these high S/Ns on LRISp. We explore high S/N techniques such as defocusing and slit-stepping during integration with high spectral and spatial oversampling. We find that the instrument flexure and interference fringes introduced by the achromatic retarders create artificial signals at 0.5% levels in the red channel which mimic real stellar signals and limit the sensitivity and calibration stability of LRISp. Careful spectral extraction and data filtering algorithms can remove these error sources. For faint targets and long exposures, cosmic ray hits are frequent and present a major limitation to the upgraded deep depletion red-channel CCD. These must be corrected to the same high S/N levels, requiring careful spectral extraction using iterative filtering algorithms. We demonstrate here characterization of these sources of instrumental polarization artifacts and present several methods used to successfully overcome these limitations. We have measured the linear to circular cross-talk and find it to be roughly 5%, consistent with the known instrument limitations. We show spectropolarimetric signals on brown dwarfs are clearly detectable at 0.2% amplitudes with sensitivities better than 0.05% at full spectral sampling in atomic and molecular bands. Future LRISp users can perform high-sensitivity observations with high-quality calibration when following the described algorithms.

CASSIS: THE CORNELL ATLAS OF SPITZER /INFRARED SPECTROGRAPH SOURCES. II. HIGH-RESOLUTION OBSERVATIONS

The Infrared Spectrograph (IRS) on board the Spitzer Space Telescope observed about 15,000 objects during the cryogenic mission lifetime. Observations provided low-resolution (R~60-127) spectra over ~5-38um and high-resolution (R~600) spectra over ~10-37um. The Cornell Atlas of Spitzer/IRS Sources (CASSIS) was created to provide publishable quality spectra to the community. Low-resolution spectra have been available in CASSIS since 2011, and we present here the addition of the high-resolution spectra. The high-resolution observations represent approximately one third of all staring observations performed with the IRS instrument. While low-resolution observations are adapted to faint objects and/or broad spectral features (e.g., dust continuum, molecular bands), high-resolution observations allow more accurate measurements of narrow features (e.g., ionic emission lines) as well as a better sampling of the spectral profile of various features. Given the narrow aperture of the two high-resolution modules, cosmic ray hits and spurious features usually plague the spectra. Our pipeline is designed to minimize these effects through various improvements. A super sampled point-spread function was created in order to enable the optimal extraction in addition to the full aperture extraction. The pipeline selects the best extraction method based on the spatial extent of the object. For unresolved sources, the optimal extraction provides a significant improvement in signal-to-noise ratio over a full aperture extraction. We have developed several techniques for optimal extraction, including a differential method that eliminates low-level rogue pixels (even when no dedicated background observation was performed). The updated CASSIS repository now includes all the spectra ever taken by the IRS, with the exception of mapping observations.

The Detection of Outliers in Nondestructive Integrations with the Generalized Extreme Studentized Deviate Test

The identification of outliers in nondestructive integrations represents an important step in the routine processing of on-orbit or ground observations and in providing data of the highest quality to the observer. We present the Generalized Extreme Studentized Deviate (ESD) Test for multiple outliers and apply it to the identification of cosmic ray hits in simulated integrations and in ground laboratory observations collected with a James Webb Space Telescope (JWST) near-IR detector. The completeness ratio and fraction of false detections of the simulated cosmic-ray impacts are characterized for a range of parameters expected in real-world observations. Random Telegraph Noise (RTN) pixels, which are intrinsically variable, are also successfully identified with this test. The algorithm of this versatile statistical test is straightforward to implement and can serve to flag and mask bad or variable pixels in automated data calibration pipelines.

Scanamorphos: A Map-making Software for Herschel and Similar Scanning Bolometer Arrays

Scanamorphos is public software available to postprocess scan observations performed with the Herschel photometer arrays. This postprocessing mainly consists in subtracting the total low-frequency noise (both its thermal and nonthermal components), masking high-frequency artefacts such as cosmic ray hits, and projecting the data onto a map. Although it was developed for Herschel, it is also applicable with minimal adjustment to scan observations made with some other imaging arrays subjected to low-frequency noise, provided they entail sufficient redundancy; it was successfully applied to P-Artémis, an instrument operating on the APEX telescope. Contrary to matrix-inversion softwares and high-pass filters, Scanamorphos does not assume any particular noise model, and does not apply any Fourier-space filtering to the data, but is an empirical tool using purely the redundancy built in the observations—taking advantage of the fact that each portion of the sky is sampled at multiple times by multiple bolometers. It is an interactive software in the sense that the user is allowed to optionally visualize and control results at each intermediate step, but the processing is fully automated. This paper describes the principles and algorithm of Scanamorphos and presents several examples of application.

The North Ecliptic Pole Wide survey of AKARI: a near- and mid-infrared source catalog

We present a photometric catalog of infrared (IR) sources based on the North Ecliptic PoleWide field (NEP-Wide) survey of AKARI, which is an infrared space telescope launched by Japan. The NEP-Wide survey covered 5.4 deg2 area, a nearly circular shape centered on the North Ecliptic Pole, using nine photometric filter-bands from 2 - 25 {\mu}m of the Infrared Camera (IRC). Extensive efforts were made to reduce possible false objects due to cosmic ray hits, multiplexer bleeding phenomena around bright sources, and other artifacts. The number of detected sources varied depending on the filter band: with about 109,000 sources being cataloged in the near-IR bands at 2 - 5 {\mu}m, about 20,000 sources in the shorter parts of the mid-IR bands between 7 - 11 {\mu}m, and about 16,000 sources in the longer parts of the mid-IR bands, with \sim 4,000 sources at 24 {\mu}m. The estimated 5? detection limits are approximately 21 magnitude (mag) in the 2 - 5 {\mu}m bands, 19.5 - 19 mag in the 7 - 11 {\mu}m, and 18.8 - 18.5 mag in the 15 - 24 {\mu}m bands in the AB magnitude scale. The completenesses for those bands were evaluated as a function of magnitude: the 50% completeness limits are about 19.8 mag at 3 {\mu}m, 18.6 mag at 9 {\mu}m, and 18 mag at 18 {\mu}m band, respectively. To construct a reliable source catalog, all of the detected sources were examined by matching them with those in other wavelength data, including optical and ground-based near-IR bands. The final band-merged catalog contains about 114,800 sources detected in the IRC filter bands. The properties of the sources are presented in terms of the distributions in various color-color diagrams.

PyCosmic: a robust method to detect cosmics in CALIFA and other fiber-fed integral-field spectroscopy datasets

[Abridged] Detecting cosmic ray hits (cosmics) in fiber-fed IFS data of single exposures is a challenging task, because of the complex signal recorded by IFS instruments. Existing detection algorithms are commonly found to be unreliable in the case of IFS data and the optimal parameter settings are usually unknown a-priori for a given dataset. The CALIFA survey generates hundreds of IFS datasets for which a reliable and robust detection algorithm for cosmics is required as an important part of the fully automatic CALIFA data reduction pipeline. We developed a novel algorithm, PyCosmic, which combines the edge-detection algorithm of L.A.Cosmic with a point-spread function convolution scheme. We generated mock data to compute the efficiency of different algorithms for a wide range of characteristic fibre-fed IFS datasets using the PMAS and VIMOS IFS instruments as representative cases. PyCosmic is the only algorithm that achieves an acceptable detection performance for CALIFA data. We find that PyCosmic is the most robust tool with a detection rate of >~90% and a false detection rate <5% for any of the tested IFS data. It has one less free parameter than the L.A.Cosmic algorithm. Only for strongly undersampled IFS data does L.A.Cosmic exceed the performance of PyCosmic by a few per cent. DCR never reaches the efficiency of the other two algorithms and should only be used if computational speed is a concern. Thus, PyCosmic appears to be the most versatile cosmics detection algorithm for IFS data. It is implemented in the new CALIFA data reduction pipeline as well as in recent versions of the multi-instrument IFS pipeline P3D.

Search for satellites near (21) Lutetia using OSIRIS/Rosetta images

On 2010 July 10 the ESA Rosetta mission flew by the large asteroid (21) Lutetia. One of the scientific goals of the onboard OSIRIS instrument was the search for satellites of the asteroid, with more than 20 images specifically dedicated to this topic. An observational campaign was devised with a selection of filters and exposure times tailored to maximize the possibility of detecting small companions and determining their bound orbits. Data were analyzed with suitable methods to remove cosmic ray hits and known background objects, in order to search for persistent detections of potential interesting flux sources. We found no unambiguous detections of a satellite larger than ∼160m inside the entire sphere of gravitational influence. Our search confirmed the absence of bound companions larger than ∼30m inside 20 primary radii. These limits are a factor of ∼30 smaller than the values reported so far from large ground-based telescopes using adaptive optics and from the Hubble Space Telescope.

Cosmic-ray rejection for single spectroscopic CCD images by means of template matching

This article describes a novel algorithm for cosmic-ray rejection in single spectroscopic CCD images. This algorithm is based on a variation of template matching. It focuses on identifying those pixels belong to spectra, while other conventional algorithms tried to locate the cosmic-ray hits directly. The main principle is applying template matching to find suspicious blocks, which is followed by surface patching to locate the legitimate pixels accurately. Therefore, the rest pixels are the ones corrupted by cosmic-ray hits. Meanwhile, almost all the parameters are automatically extracted from the images. Examples of its performance are given for both simulated and observed images. It shows an advantage of significantly low false alarm rate with relatively high detection rate (© 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim)

Cosmic-Ray Detection Based on Gray-Scale Morphology of Spectroscopic CCD Images

AbstractDetection of cosmic-ray hits is important because of their disturbing effects on image processing. An algorithm based on gray-scale morphology to detect cosmic-rays in single spectroscopic CCD images is presented in this paper. Based on the morphological differences between cosmic-ray hits and spectra, the operations of erosion and dilation are used to deal with images successively, with appropriate structuring elements. Points of cosmic-ray hits can be detected by analyzing the variation between data pre- and post-operation. Examples of the algorithm are given at the end of this paper, with several groups of both simulated LAMOST images and observed SDSS images to show its performance.