Subband Matrix Information Geometry Detector in Heterogeneous Clutter

Abstract

Matrix information geometry (MIG) detector, which converts sample data to Hermitian positive definite (HPD) matrices located on HPD matrix manifold, provides an innovative scheme for target detection. In this paper, a subband MIG detector is proposed to detect target submerged into heterogeneous clutter background with short pulses. More precisely, subband filtering is firstly performed to suppress strong clutter by designing a discrete Fourier transform modulated filter bank. Then, a set of HPD matrices are modeled by the filtered data and a HPD matrix manifold is formed. In each subband, the detection from geometric consideration on the manifold is derived. Thus, a subband MIG detector is formulated and information divergence is utilized to measure the dissimilarity of the observed data and the clutter. Finally, numerical experiments based on simulated data and real sea clutter data show that the proposed method can achieve better detection performance compared with the traditional methods.