COHERENT RADAR DETECTION AGAINST K-DISTRIBUTED CLUTTER WITH PARTIALLYCORRELATED TEXTURE

This paper deals with the radar detection of targets embedded in K-dis tributed clutter with partially correlated texture. A model for K-dist ributed clutter is first proposed, which is consistent with the main m odels of non-Gaussian scattering, and its generality is assessed. The optimum coherent detection of targets against such a clutter is then s tudied, a parallel implementation and a sequential implementation of t he optimum detector are given. The optimum detector can be interpreted as the conditional estimate of the Gaussian likelihood ratio of the s peckle term averaged over the correlated texture distribution. On this basis a suboptimum detection scheme is derived where the conditional mean estimation is replaced with a maximum a posteriori estimate. A re cursive implementation is provided, which exploits the correlation pro perties of both texture and speckle components.