Modeling And Targei Detection In Sea Clutter

Detection in sea clutter is a very important issue. However， the design andperformance prediction of radar detectors in sea environments are among the mostcomplex problems in radar signal processing. Since sea clutter is highly dependent onthe ocean state， radar grazing angle， wind velocity and direction， it has strongnonlinear dynamic characteristics. In range domain， sea clutter features as a lot ofspikes， which will decrease the target detection performance， especially for the targetof low speed and low RCS. However， it is widely used in military affairs and maritimerescue to detect low-speed small target on the ocean surface. This motivates the workpresented in this paper.Presently， this paper introduces the model of scatter coefficient， the probabilitydistribute model， the power spectrum model of the sea clutter， then simulates andgets the sea clutter simulate sequences which’s probability distribute and powerspectrum are wanted by using these models.As a suitable for describing the radar clutter distribution model， Rayleighdistribution， Weibull distribution， Lognormal distribution and K-distribution fourmodel has a pivotal low. his paper discusses the distribution of several model andsimulation， also using Zero Memory NonLinearity (ZMNL)， produced with theabove model simulation of the distribution of the clutter.As the present study on sea clutter are shown， based on the analysis of sea clutterDoppler spectrum characteristics， two target detection algorithms are proposed，namely， a Short Average Power and a Doppler Centrod. The detection performances ofthe proposed algorithms are evaluated based on the data collected by the McMasterIPIX radar at the east coast of Canada. It is noted that the two algorithms have goodsupplement effect. The two features and the variance in time domain can be used inSVM classifier， and we have also achieved good results.