Small Target Detection Algorithm Research In Sea Clutter Based On Spatio-Temporal Chaos

Small target detection is widely used in military defense, maritime trade, maritime rescue and other fields. Sea clutter is different from ground clutter, which is a distribution of scattering phenomenon and shows more dynamic characteristics, therefore strong sea clutter often becomes the main interference in small weak signal. How to deal with sea clutter will directly affect the capability of target detection in the marine environment.Sea clutter intensity distribution primarily depends on marine power, solar elevation and location of stations, and the nonlinear ocean dynamic results in the nonlinear characteristics of sea clutter image sequence. Based on the chaos theory of nonlinear dynamic, the characteristics of sea clutter image sequence are analyzed, and the space-time chaos theory is used to suppress sea clutter. This paper mainly researches on the detection of the small targets of the sea clutter. The main research results are as follows:(1) The theory of chaotic dynamics includes the phase space reconstruction and chaotic feature recognition. Both calculating chaotic characteristics and predicting chaos model should be carried out in the phase space, so phase space reconstruction is the indispensable link for processing nonlinear data sequence. In order to calculate the characteristic parameters of sea clutter intensity sequence, G-P(Grassberger-Procaccia) method is chosen to calculate the correlation dimension and small data sets is used to calculate the Lyapunov exponent. Experimental results show that sea clutter intensity data sequence of images have the finite correlation dimension and a positive largest Lyapunov exponent. Therefore, these results show that the sequence of sea clutter images is chaos.(2) Considering spatial region and temporal region, prediction method based on spatial and temporal chaotic sequence for sea clutter is proposed in this paper. Firstly, the data series is selected to verify chaotic characteristics: for selecting spatial intensity data series, the moving direction of waves must be determined, and a continuous gray values in this direction is selected as a sequence of intensity data for chaos prediction; for selecting temporal intensity data series, the gray values at the same position on a continuous frame image are regarded as a group of sequence. Secondly, neural networks is used to predict the chaotic spatial data series and temporal data series. Based on the phase space reconstruction theory, RBF neural network reconstructs dynamic model of sea clutter.(3) A method based on the space-time chaos of image sequences is proposed to detect small target in sea clutter. The method fully reflects the spatial-temporal motion information of sea clutter by using a coupled map lattice. RBF neural network is used to reconstruct dynamic model of sea clutter, and it is applied to predict and cancel sea clutter. Finally small targets can be detected from the sea clutter background. In this paper, Signal Noise Ratio(SNR) is selected to evaluate the experimental results. The performance of spatial, temporal and spatial-temporal detection are compared, it is concluded: small target detection in sea clutter by using spatio-temporal chaos can suppress sea clutter preferably and separate target from sea clutter background. Consequently, small target detection in sea clutter by using spatio-temporal chaos can effectively detect the useful target signal overwhelmed by the sea clutter.Theoretical analysis and experimental results show that small target detection in sea clutter by using spatio-temporal chaos in this paper has the advantage of suppressing sea clutter, and it can effectively improve the target detection performance. So small target detection in sea clutter based on spatio-temporal chaos has a certain theoretical significance and practical value.