We present a generalized non-negative factorization (NMF)-based data reduction pipeline for circumstellar disk and exoplanet detection. By using an adaptable pre-processing routine that applies algorithmic masks and corrections to improper data, we are able to easily offload the computationally-intensive NMF algorithm to a Graphics Processing Unit (GPU), significantly increasing computational efficiency. NMF has been shown to better preserve disk structural features compared to other post-processing approaches and has demonstrated improvements in the analysis of archival data. The adaptive pre-processing routine of this pipeline, which automatically aligns and applies image corrections to the raw data, is shown to significantly improve chromatic halo suppression. Utilizing HST-STIS and JWST-MIRI coronagraphic datasets, we demonstrate a factor of five increase in real-time computational efficiency by using GPUs to perform NMF compared to using CPUs. Additionally, we demonstrate the usefulness of higher numbers of NMF components with SNR and contrast improvements, which necessitates the use of a more computationally efficient approach for data reduction.
Because of bright starlight leakage in coronagraphic raw images, faint astrophysical objects such as exoplanets can only be detected using powerful point spread function (PSF) subtraction algorithms. However, these algorithms have strong effects on faint objects of interest, and often prevent precise spectroscopic analysis and scattering property measurements of circumstellar disks. For this reason, PSF-subtraction effects is currently the main limitations to the precise characterization of exoplanetary dust with scattered-light imaging. Forward modeling techniques have long been developed for point source objects (Pueyo 2016). However, Forward Modeling with disks is complicated by the fact that the disk cannot be simplified using a simple point source convolved by the PSF as the astrophysical model; all hypothetical disk morphologies must be explored to understand the subtle and non-linear effects of the PSF subtraction algorithm on the shape and local geometry of these systems. Because of their complex geometries, the forward-modeling process has to be repeated tens or hundred of thousands of times on disks with slightly different physical properties. All of these geometries are then compared to the PSF-subtracted image of the data, within an MCMC or a Chi-square wrapper. In this paper, we present here DiskFM, a new open-source algorithm included in the PSF subtraction algo- rithms package pyKLIP. This code allows to produce fast forward-modeling for a variety of observation strategies (ADI, SDI, ADI+SDI, RDI). pyKLIP has already been used for SPHERE/IRDIS and GPI data. It is readily available on all instruments supported by pyKLIP (SPHERE/IFS, SCExAO/CHARIS), and can be quickly adapted for other coronagraphic instruments.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.