Abstract: The high resolution radar target echoes can be referred to as a distributed target. In the non-Gaussian background, the adaptive detection of a distributed target is investigated without secondary data. Firstly, the set of target scatterers are estimated roughly by means of the order statistics detection theory and the recursive estimator of covariance matrix. Then, an approximately maximum likelihood estimate of the covariance matrix is acquired by the recursive estimator. Furthermore, an adaptive detector without secondary data(ADWSD)is derived for the distributed target. The proposed detector is almost of constant false alarm rate for a non-Gaussian clutter. In addition, the ADWSD performs much better than the existing detector of modified generalized likelihood ratio test. Finally, the simulation results show that, as the clutter spikiness increases, the detection performance of ADWSD is improved, especially for the low signal-to-clutter ratio scenarios; and it is also robust with respect to different correlations of clutter.

Key words: non-Gaussian clutter, distributed target detection, modified generalized likelihood ratio test, maximum likelihood, covariance matrix