Similarity‐based multipath suppression algorithm for through‐wall imaging radar

Abstract

This study deals with the multipaths suppression problem in a shielded environment with a through-wall imaging radar. The authors propose a similarity-based algorithm to eliminate the multipath ghosts for the scenario where some targets overlap with other targets’ multipath ghosts. Specifically, the authors first extract the connected domains of the radar image after binarisation, and then put them into a vector in ascending order based on the distances between their average coordinates and the array centre. Second, they divide the vector into two sub-vectors and calculate the distances between each sub-vector and all sensor locations along different paths. Third, they construct the similarity matrix using these two sub-vectors and further derive the matrix elements larger than the similarity threshold. Finally, decision rules are designed to eliminate the multipath ghosts without losing true targets. At the analyse stage, the authors assess the effectiveness of the proposed approach through simulation results showing its ability of well removing multipath ghosts.