Bayesian statistics approach to imaging of aperture synthesis data: RESOLVE meets ALMA

TL;DR

This paper introduces the RESOLVE Bayesian inference algorithm for imaging ALMA data, offering an alternative to the traditional CLEAN method, especially for extended structures and negative fluxes.

Contribution

It presents the application of the RESOLVE Bayesian imaging algorithm to ALMA data, improving image reconstruction for complex and extended astronomical sources.

Findings

RESOLVE outperforms CLEAN in reconstructing extended structures.

RESOLVE handles negative fluxes more physically.

Comparison with actual ALMA data demonstrates improved image quality.

Abstract

The Atacama Large Millimeter/submillimeter Array (ALMA) is currently revolutionizing observational astrophysics. The aperture synthesis technique provides angular resolution otherwise unachievable with the conventional single-aperture telescope. However, recovering the image from the inherently undersampled data is a challenging task. The CLEAN algorithm has proven successful and reliable and is commonly used in imaging the interferometric observations. It is not, however, free of limitations. Point-source assumption, central to the CLEAN is not optimal for the extended structures of molecular gas recovered by ALMA. Additionally, negative fluxes recovered with CLEAN are not physical. This begs to search for alternatives that would be better suited for specific science cases. We present the recent developments in imaging ALMA data using Bayesian inference techniques, namely the RESOLVE algorithm This algorithm, based on information field theory \cite{Ensslin2013}, has been already successfully applied to image the Very Large Array data. We compare the capability of both CLEAN and RESOLVE to recover known sky signal, convoluted with the simulator of ALMA observation data and we investigate the problem with a set of actual ALMA observations.