Synthesis and analysis of detection algorithm of distributed sources against background with distributed clutter

The paper deals with an algorithm for detection of a broadband signal received from a multiple-unit source against the background with architecturally-like clutters that differ by location and additive space-uncorrelated noise. It is shown that the optimal detection algorithm represents the weight processing in the "ray space" of the transformed vector of observed data and does not demand access of any large matrix. A maximum likelihood ratio is chosen as the optimality criterion. After determining the inverse of the correlation function, expression for the optimal weight function is obtained. The design ratios for determination of detection characteristics under the Gaussian statistics of signals and clutters are given. The results confirm the property of the suggested algorithm: at low signal/noise ratios, efficiency of point targets detection is better, and at the large ratios the distributed targets are better detected.