A Decentralized Framework for Radio-interferometric Image Reconstruction

TL;DR

This paper introduces a decentralized, parallelized framework for radio-interferometric image reconstruction that significantly accelerates processing times for large datasets while maintaining image quality.

Contribution

It proposes a novel decentralized approach that parallelizes image reconstruction by spatial frequency, enabling scalable processing for large radio telescope data.

Findings

Achieves near 2x speedup on large datasets

Maintains comparable image quality to serial methods

Applicable to multiscale CLEAN and sparsity regularized reconstructions

Abstract

The advent of large aperture arrays, such as the ones currently under construction for the SKA project, allows for observing the Universe in the radio-spectrum at unprecedented resolution and sensitivity. To process the enormous amounts of data produced by these telescopes, scalable software pipelines are required. This paper helps address this by proposing a framework that allows for decentralized radio-interferometric image reconstruction, parallelizing by spatial frequency. This is achieved by creating pseudo-full-resolution problems for each node by using the local visibilities together with previous major cycle reconstructed images from the other nodes. We apply the proposed framework to both multiscale CLEAN and sparsity regularized convex reconstruction and compare them to their serial counterparts across four different data sets of varying properties in the context of two visibility partitions. We found that the parallelization framework allows for significantly improved reconstruction times for images of similar quality. This was especially the case for our larger datasets where we were able to achieve close to the optimal $2\times$ speedup.