Automatic censored CFAR detection for nonhomogeneous environments

In constant false alarm rate (CFAR) detection, the clutter background in the cell under test is estimated by combining the outputs of the nearby resolution cells. In a nonhomogeneous background environment, the false alarm regulation capabilities and the detection performance of a CFAR detector depend on the robustness of the noise level estimator. The authors propose and study automatic censoring algorithms for nonhomogeneous clutter background environments. A cell-by-cell criterion for accepting or rejecting reference samples is introduced. The detection schemes that are considered do not require any a priori knowledge about the background environment. The case where a clutter edge is present in the reference window is considered, and the case of multiple target situations is studied. The case when both interfering targets and a clutter edge appear in the reference window of the cell under test is considered. The performances of the proposed schemes are compared with that of the trimmed-mean (TM) CFAR detector, in which the censoring points are preassigned