Small Target Detection Based on Noncoherent Radial Velocity Spectrum of High-Resolution Sea Clutter

Abstract

This paper proposes a multiscan joint non-coherent detection method of small targets in high-resolution sea clutter with spatial-temporally structural textures induced by large-scale swells and waves, which is composed of intrascan non-coherent integration at each spatial resolution cell followed by interscan integration on a radial velocity template. The main contributions of the paper are as follows. Firstly, a concept of non-coherent radial velocity spectrum (NCRVS) is proposed to model the non-homogeneity of multiscan integration of sea clutter with structural textures. In the NCRVS, the non-coherent intrascan integration and the interscan integration by passing the retrospective filter (RF) bank are fully combined. Secondly, the NCRVS of sea clutter has different statistics at individual radial velocity bins and the lognormal distributions are used to fit their statistics and an outlier-robust analytical bi-percentile estimator is constructed for parameter estimation. Thirdly, based on the lognormal-distributed NCRVS model, a double-channel multiscan joint detector (DC-MJD) is proposed, where the intrascan non-coherent integration channel is used to find floating or low radial velocity targets and the whitening plus intrascan non-coherent integration channel is used to find target with high radial velocity. Simulated data and measured X-band radar data are used to verify the DC-MJD detector. In comparison with the existing three multiscan detectors, the DC-MJD detector improves the detection probability by 96% in the measured data with a test target.