Lean CLEAN: Deconvolution Algorithm for Radar Imaging of Contiguous Targets

Abstract

High-resolution microwave images of contiguous targets often suffer from sidelobe artifacts. Deconvolution algorithms, such as Coherent CLEAN, do not perform well in these scenarios because of spurious peaks and peak mislocations. These spurious and mislocated peaks are often interpreted as real targets by the CLEAN algorithm. In this paper, we propose the Lean CLEAN technique that addresses this problem. The Lean CLEAN technique consists of a postprocessing block that works with the CLEAN algorithm. It detects and discards the spurious peaks and then consolidates the multiple small peaks corresponding to mislocated targets. Extensive simulations suggest that the proposed technique works well with different CLEAN techniques.