Analysis of non-Coherent CFAR Detection in CG-LNT Radar Clutter

This communication deals with the problem of non-coherent CFAR (Constant False Alarm Rate) detection in homogeneous and heterogeneous CG-LNT (Compound Gaussian-Lognormal Texture) clutter. The distribution of the latter with and without thermal noise is presented and possesses a strong ability to model various sea-clutter scenes. The performances of CA (Cell Averaging)-, SO (Smallest Of)-, GO (Greatest Of)-and OS (Order Statistic)-CFAR algorithms are investigated for a known standard deviation parameter. In the presence of secondary targets WH (Weber Hykin)- and WHOS-CFAR detectors are also examined where their CFAR properties are shown to be better than those obtained by OS-CFAR procedure.