Range And Doppler Distributed Targets Detection In Sea Clutter

When a high range resolution radar is used to detect surface targets, the probability density function (pdf) of sea clutter is no longer Rayleigh form due to the fact that the area of the surface illuminated of every single range bin is small enough so that the central-limit theorem cannot be satisfied. Meanwhile, because range resolution is high enough, radar can identify the structure of sea surface, so the clutter is highly correlated in time and space. The detectors which are designed based on Rayleigh clutter model will not work well on high range resolution radar. This paper aims to study how to design a detector for high range resolution radar using Non-Gaussian clutter model.First, since there exist no closed-form analytical probability distributions in many Non-Gaussian correlated clutters, we study how to generate simulation sea clutters with specified pdf and correlation attributes. SSE/SSE2 instruction set is used in order to increase the speed of simulation clutter generation.Second, we study the noncoherent slow target detector (NSTAD) algorithm for slow moving targets. This algorithm is designed and tested using an ideal clutter model. When we use simulation clutter to test this algorithm, it is nearly as good as LRT, and this algorithm will perform much better than likelihood ratio test (LRT) under real sea clutter.Then we study the generalize likelihood ratio test GLRT subspace detector for range and Doppler distributed targets.Finally, a sea clutter suppression algorithm based on PCA is used to further improve the performance of detector.Simulation results show the effectiveness of the detectors.