Research On Target Self-adaptation Detection Algorithm In Non-stationary Sea Clutter Environment

The problem of radar target detection in sea clutter is always a hot and difficult research topic in the radar signal processing field.With the increasing of radar range resolution and the decreasing of incident residual angle,sea clutter shows significant non-Gaussian and non-stationary characteristics in both time and space dimensions.The detection performance of the traditional adaptive coherent detectors in non-stationary sea clutter is limited by the interference of the strong sea clutter and the limited coherent integration time.To overcome the limitations from the non-Gaussian and nonstationary characteristics,two different adaptive coherent detectors are presented: one is a smooth generalized likelihood ratio test(S-GLRT)detector for the partially homogeneous sea clutter environment,and the other is subband segmented adaptive normalized matched filter(SS-ANMF)detector for spatial-temporal non-stationary sea clutter environment.In order to alleviate the large transient fluctuation of the partially homogeneous sea clutter,a smoothing factor is used in the SGLRT to control the instantaneous power of sea clutter.The analytic expression of probability of false alarm is derived to verify the constant false alarm ratio with respect to the scale parameter of texture.Without an additional computational complexity,experimental results show that S-GLRT outperforms GLRT in the partially homogeneous sea clutter environment.The SS-ANMF is composed by a linear phase discrete Fourier transform(DFT)modulated filter bank and a segmented ANMF detector in each subband.Resorted to the subband decomposition in the frequencydomain,the clutter in and out of band is restricted and the short-term stationarity of clutter is enhanced.Besides,by summing over the maximum responses of segmented ANMF of all subbands within a cumulative time,the restriction from the sample size is lifted and the accumulated time is extended.The experimental results using real data show that the SS-ANMF detector outperforms the compared ones.