Analysis Of Sea Clutter Characteristics And Processing Of Target Detection

Analysis of sea clutter characteristic and target detection is one of research branch of Radar signal processing fields. How to realize accurate modeling of sea clutter and how to detect targets from sea clutter is a very important research direction both in civil and military fields. Based on statistical modeling, chaotic and nonlinear dynamic modeling and AR modeling for IPIX radar real sea clutters, some research and analysis is conducted in this paper.The statistical characteristic of the IPIX radar sea clutter is studied in this paper. By analyzing the calculation and select of statistical parameters, the amplitude of the sea clutter were fitted with several statistical distributions. Since the statistical characteristic is varied subject to different polarization and sea state, it is difficult to accurately establish the statistical model of sea clutter. In another word, it is hard to establish an accurate model by using a single traditional statistical distribution model. Then it is proposed to treat the sea clutter as a stationary AR model and establish AR model by estimate their order and parameters. Accordingly, a linear predictor and detector which can accurately simulate the sea clutter are established.Based on the nonlinear dynamic characteristic of the sea clutter and phase space reconstruction theory, nonlinear phase space of IPIX real sea clutter is reconstructed. For that suppose, three methods for calculating delay timeτ(Autocorrelation function method, mutual information method, C-C method and its improved algorithm) are discussed in this paper. Besides, two methods for determining embedding dimension m (false nearest neighbor and Cao method) are concerned. For another nonlinear prediction phase space reconstruction method, determining method for coefficients of the Volterra series filters is analyzed. In order to accurately recover and predict the sea clutter series, the selection on parameters of phase space reconstruction shall be as precise as possible. Similarly, two nonlinear prediction predictors and detectors are established based on AR model and Radial basis function network (RBFN) and Volterra series filter network (VSFN). The established linear and nonlinear predictor and detector based separately on AR model and RBFN and VSFN were applied to the same IPIX radar real sea clutter data for prediction. Compared with the prediction beneath simulated chaotic characteristic sea clutter background, a target detection method based on this prediction model is presented.Micro-Doppler effect existed in non-coherent radar Intermediate Frequency is concerned in this paper. By analyzing its spectrum characteristic and time-frequency characteristic, it can be found that it is difficult to distinguish this micro-Doppler shift only by frequency transform and time-frequency analysis method. A target detection method based on fuzzy C-mean clustering algorithm is proposed, which can realize the signal detection in time domain. Further, noise effect for the characteristic analysis and target detection algorithm is discussed in this paper. The algorithm based on fuzzy clustering is a new attempt in non-coherent radar intermediate frequency signal processing. So the theme of this article has both practical and theoretical values.