Parameterized CLEAN deconvolution in radio synthesis imaging

Abstract

This paper reviews parameterized CLEAN deconvolution, which is widely used in radio synthesis imaging to remove the effects of sidelobes from the point-spread function caused by incomplete sampling by the radio telescope array. At the same time, different forms of parameterization and components are provided, as well as methods for approximating the true sky brightness. In recent years, a large number of variants of the CLEAN algorithm have been proposed to deliver faster and better reconstruction of extended emission. The diversity of algorithms has stemmed from the need to deal with different situations as well as optimizing the previous algorithms. In this paper, these CLEAN deconvolution algorithms are classified as scale-free, multi-scale and adaptive-scale deconvolution algorithms based on their different sky-parameterization methods. In general, scale-free algorithms are more efficient when dealing with compact sources, while multi-scale and adaptive-scale algorithms are more efficient when handing extended sources. We will cover the details of these algorithms, such as how they handle the background, their parameterization and the differences between them. In particular, we discuss the latest algorithm, which has been able to efficiently handle both compact and extended sources simultaneously via the deep integration of scale-free and adaptive-scale algorithms. We also mentioned recent developments in other important deconvolution methods and compared them with CLEAN deconvolution.