Source Detection with Interferometric Datasets

Abstract

The detection of sources in interferometric radio data typically relies on extracting information from images, formed by Fourier transform of the underlying visibility dataset, and CLEANed of contaminating sidelobes through iterative deconvolution. Variable and transient radio sources span a large range of variability timescales, and their study has the potential to enhance our knowledge of the dynamic universe. Their detection and classification involve large data rates and non-stationary PSFs, commensal observing programs and ambitious science goals, and will demand a paradigm shift in the deployment of next-generation instruments. Optimal source detection and classification in real time requires efficient and automated algorithms. On short time-scales variability can be probed with an optimal matched filter detector applied directly to the visibility dataset. This paper shows the design of such a detector, and some preliminary detection performance results.