Adaptive detection of distributed target in the presence of signal mismatch in compound Gaussian clutter

Abstract This paper deals with adaptive detection of distributed targets in the presence of signal mismatch in compound Gaussian clutter. The texture is assumed random to match the non-Gaussian environment. To increase the capability of rejecting mismatched signals, a fictitious subspace signal is introduced under the null hypothesis. Under the above assumptions, we propose two adaptive selective detectors by using two-step generalized likelihood ratio test (GLRT) and maximum a-posteriori (MAP) GLRT design criteria. The two new detectors are demonstrated to achieve constant false alarm rate property with respect to the speckle covariance matrix. Numerical simulations indicate that the new detectors outperform the conventional one in terms of selectivity property.