Search for: All records

Rolling shutter distortion is highly undesirable for photography and computer vision algorithms (e.g., visual SLAM) because pixels can be potentially captured at different times and poses. In this paper, we propose a deep neural network to predict depth and row-wise pose from a single image for rolling shutter correction. Our contribution in this work is to incorporate inertial measurement unit (IMU) data into the pose refinement process, which, compared to the state-of-the-art, greatly enhances the pose prediction. The improved accuracy and robustness make it possible for numerous vision algorithms to use imagery captured by rolling shutter cameras and produce highly accurate results. We also extend a dataset to have real rolling shutter images, IMU data, depth maps, camera poses, and corresponding global shutter images for rolling shutter correction training. We demonstrate the efficacy of the proposed method by evaluating the performance of Direct Sparse Odometry (DSO) algorithm on rolling shutter imagery corrected using the proposed approach. Results show marked improvements of the DSO algorithm over using uncorrected imagery, validating the proposed approach. 

We present an observational study of the luminous red nova (LRN) AT 2021biy in the nearby galaxy NGC 4631. The field of the object was routinely imaged during the pre-eruptive stage by synoptic surveys, but the transient was detected only at a few epochs from ∼231 days before maximum brightness. The LRN outburst was monitored with unprecedented cadence both photometrically and spectroscopically. AT 2021biy shows a short-duration blue peak, with a bolometric luminosity of ∼1.6 × 10 41 erg s −1 , followed by the longest plateau among LRNe to date, with a duration of 210 days. A late-time hump in the light curve was also observed, possibly produced by a shell-shell collision. AT 2021biy exhibits the typical spectral evolution of LRNe. Early-time spectra are characterised by a blue continuum and prominent H emission lines. Then, the continuum becomes redder, resembling that of a K-type star with a forest of metal absorption lines during the plateau phase. Finally, late-time spectra show a very red continuum ( T BB  ≈ 2050 K) with molecular features (e.g., TiO) resembling those of M-type stars. Spectropolarimetric analysis indicates that AT 2021biy has local dust properties similar to those of V838 Mon in the Milky Way Galaxy. Inspection of archival Hubble Space Telescope data taken on 2003 August 3 reveals a ∼20 M ⊙ progenitor candidate with log ( L / L ⊙ ) = 5.0 dex and T eff  = 5900 K at solar metallicity. The above luminosity and colour match those of a luminous yellow supergiant. Most likely, this source is a close binary, with a 17–24 M ⊙ primary component. 

In this paper we report the results of the first ~four years of spectroscopic and photometric monitoring of the Type IIn supernova SN 2015da (also known as PSN J13522411+3941286, or iPTF16tu). The supernova exploded in the nearby spiral galaxy NGC 5337 in a relatively highly extinguished environment. The transient showed prominent narrow Balmer lines in emission at all times and a slow rise to maximum in all bands. In addition, early observations performed by amateur astronomers give a very well-constrained explosion epoch. The observables are consistent with continuous interaction between the supernova ejecta and a dense and extended H-rich circumstellar medium. The presence of such an extended and dense medium is difficult to reconcile with standard stellar evolution models, since the metallicity at the position of SN 2015da seems to be slightly subsolar. Interaction is likely the mechanism powering the light curve, as confirmed by the analysis of the pseudo bolometric light curve, which gives a total radiated energy ≳ 10 51 erg. Modeling the light curve in the context of a supernova shock breakout through a dense circumstellar medium allowed us to infer the mass of the prexisting gas to be ≃ 8 M ⊙ , with an extreme mass-loss rate for the progenitor star ≃0.6 M ⊙ yr −1 , suggesting that most of the circumstellar gas was produced during multiple eruptive events. Near- and mid-infrared observations reveal a fluxexcess in these domains, similar to those observed in SN 2010jl and other interacting transients, likely due to preexisting radiatively heated dust surrounding the supernova. By modeling the infrared excess, we infer a mass ≳ 0.4 × 10 −3 M ⊙ for the dust. 

ABSTRACT GRANDMA (Global Rapid Advanced Network Devoted to the Multi-messenger Addicts) is a network of 25 telescopes of different sizes, including both photometric and spectroscopic facilities. The network aims to coordinate follow-up observations of gravitational-wave (GW) candidate alerts, especially those with large localization uncertainties, to reduce the delay between the initial detection and the optical confirmation. In this paper, we detail GRANDMA’s observational performance during Advanced LIGO/Advanced Virgo Observing Run 3 (O3), focusing on the second part of O3; this includes summary statistics pertaining to coverage and possible astrophysical origin of the candidates. To do so, we quantify our observation efficiency in terms of delay between GW candidate trigger time, observations, and the total coverage. Using an optimized and robust coordination system, GRANDMA followed-up about 90 per cent of the GW candidate alerts, that is 49 out of 56 candidates. This led to coverage of over 9000 deg2 during O3. The delay between the GW candidate trigger and the first observation was below 1.5 h for 50 per cent of the alerts. We did not detect any electromagnetic counterparts to the GW candidates during O3, likely due to the very large localization areas (on average thousands of degrees squares) and relatively large distance of the candidates (above 200 Mpc for 60 per cent of binary neutron star, BNS candidates). We derive constraints on potential kilonova properties for two potential BNS coalescences (GW190425 and S200213t), assuming that the events’ locations were imaged.