A statistical analysis that properly characterizes sea clutter processes is indispensable both for optimum detection algorithm design and for performance prediction problems in maritime surveillance applications. In this paper, we present the statistical analysis of experimental sea clutter data collected by a high-resolution coherent monopulse radar. First, we present the amplitude statistical analyses for these clutter data. The results show that the K, Pareto, and CIG distributions can each provide good fits to the clutter data for three channels of monopulse radar. The analyses on the variations of the K distribution parameters with range suggest that the scale parameter is closely associated with the clutter powers and that the shape parameter is influenced by the sea state. Then, we focus on the correlation properties. The averaged results suggest that the temporal and spatial correlation properties are similar for the clutter of all three channels. Moreover, the clutter between the sum and difference channels is almost completely correlated in elevation and is lowly correlated in azimuth. Finally, we perform a spectral analysis, highlighting the temporal and spatial variabilities of Doppler spectra. It is found that the individual Doppler spectra in all three channels can be represented by Gaussian-shaped power spectral densities, and their centroid and width can be modeled as two separate stage linear functions of spectrum intensity.
Ship Radial Size Estimation in High-Resolution Maritime Surveillance Radars via Sparse Recovery Using Linear Programming
