As Part I of a paper series showcasing a new imaging framework, we consider the recently proposed unconstrained Sparsity Averaging Reweighted Analysis (uSARA) optimisation algorithm for wide-field, high-resolution, high-dynamic range, monochromatic intensity imaging. We reconstruct images from real radio-interferometric observations obtained with the Australian Square Kilometre Array Pathfinder (ASKAP) and present these results in comparison to the widely-used, state-of-the-art imager WSClean. Selected fields come from the ASKAP Early Science and Evolutionary Map of the Universe (EMU) Pilot surveys and contain several complex radio sources: the merging cluster system Abell 3391-95, the merging cluster SPT-CL 2023-5535, and many extended, or bent-tail, radio galaxies, including the X-shaped radio galaxy PKS 2014-558 and the ``dancing ghosts'', known collectively as PKS 2130-538. The modern framework behind uSARA utilises parallelisation and automation to solve for the w-effect and efficiently compute the measurement operator, allowing for wide-field reconstruction over the full field-of-view of individual ASKAP beams (up to 3.3 deg each). The precision capability of uSARA produces images with both super-resolution and enhanced sensitivity to diffuse components, surpassing traditional CLEAN algorithms which typically require a compromise between such yields. Our resulting monochromatic uSARA-ASKAP images of the selected data highlight both extended, diffuse emission and compact, filamentary emission at very high resolution (up to 2.2 arcsec), revealing never-before-seen structure. Here we present a validation of our uSARA-ASKAP images by comparing the morphology of reconstructed sources, measurements of diffuse flux, and spectral index maps with those obtained from images made with WSClean.
@article{arxiv.2302.14148,
title = {Scalable precision wide-field imaging in radio interferometry: I. uSARA validated on ASKAP data},
author = {Amanda G. Wilber and Arwa Dabbech and Adrian Jackson and Yves Wiaux},
journal= {arXiv preprint arXiv:2302.14148},
year = {2023}
}