CFAR binary integration detection in nonhomogeneous partially correlated clutter

The problem of binary integration detection (M-out-of-L detectors) of a Swerling II target model in nonhomogeneous partially correlated clutter is considered. The clutter is taken to be nonhomogeneous and consists of a step discontinuity. The low level contains independent thermal noise while the high level is assumed to contain spatially correlated clutter. The authors propose a new censoring procedure based on the correlation test. This procedure uses the Fisher transform to locate the clutter edge and determine whether the test cell in the clear or in the correlated clutter. The performance of M-out-of-L detectors for this binary integration detection is investigated and studied.