Radar Sea Clutter Semi-empirical Statistical Modeling And Target Detection Performance Prediction

Radar sea clutter statistical modeling and prediction of target detection performance are of great theoretical basis for the detection of small targets on the sea, the development of long-range warning radar equipped in early warning aircraft and the design of automatic detection radar system, are widely applicable in the military and civil fields. The main work is as follows:Based on the sea clutter data collected by L-band shore-based radar, the applicability of three semi-empirical sea clutter models(GIT, TSC and NRL) in littoral environments is discussed, respectively. The differences among three semi-empirical models are analyzed, and the relationship between sea clutter and wave height is studied.Four commonly used magnitude distribution models of Rayleigh, Lognormal, Weibull and K are presented, their parameter estimation methods and the determination methods of the best magnitude distribution are discussed, respectively. The best fitted models of experimental data are studied, the variation characteristics of the four distribution models parameters with the grazing angle and sea state are analyzed, respectively.The probability of false alarm for the sum of different pulses of K-distributed sea clutter and the probability of detection for different Swerling target in K-distributed sea clutter are calculated, radar range detection performance for both the three semi-empirical models and the experimental data is studied.That sea clutter have an effect on the probability of false alarm is discussed, the methods of mean level type and order statistics type constant false alarm rate(CFAR) detection in K-distributed sea clutter are studied. Three order statistics type CFAR detectors are used to detect targets in experimental sea clutter time series data, their detection performance are discussed, respectively.