Adaptive Detection in Compound-Gaussian Clutter with Partially Correlated Texture

Traditionally, the land or sea clutter with unknown covariance matrix is generally modeled as compound-Gaussian distribution with completely correlated texture in high-resolution (HR) radar. However, the detailed study on the measured clutter data from the search radar exhibits that the clutter satisfies compound-Gaussian distribution with partially correlated texture. The paper mainly addresses adaptive detection based on the two-step generalized likelihood ratio test (GLRT) criterion, and proposes the estimation methods for the unknown covariance matrix under the background of this partially correlated clutter. Numerical simulations illustrate the performance of the adaptive detectors combined with the proposed estimators under various conditions.