The Sea Echo Fractal Characteristic Analysis And Weak Target Detection

Sea echo is a scattering echo formed when radars irradiate sea surface,and also the ocean surface clutter signal,i.e.the sea clutter.Enormous interference is generated when radars detect targets on sea surface because of the sea clutter.With the increasing complexity of radar detection environment,weak targets detection in complex sea clutter background not only becomes a challenging research topic,but also is accompanied by broad application prospects.The traditional weak target detection on sea surface is based on the sea echo amplitude fitting,establishing an appropriate clutter statistical distribution model,then achieving the detection task.However,the traditional mathematical statistical model ignores the time-space variation characteristics and non-stationary of the sea echo that make the traditional model cannot describe and reflect the physical nonlinear dynamic characteristics of the ocean surface well.Also,detection algorithm based on the mathematical statistical distribution model has some limitations in practical applications.The rapid development of fractal theory,provides a new idea for sea target detection.Many fractal theory-based weak target detection algorithms have been proposed.Fractal characteristics of sea echo and the weak target detection are analyzed in this paper.Firstly,the latest fractal theory is introduced.Combined with the measured sea echo data,the fractal characteristics of sea echo are validated.The influence of target existence or not to fractal characteristics of sea echo is analyzed as well.Then Because of the single fractal characteristics of sea echo,two weak target detection algorithms are proposed.One based on box-counting dimension and the other based on the Empirical Mode Decomposition(EMD)and Hurst exponent.The Hurst exponent is theoretically proved to be reasonable test statistics.By using detection performance of single fractal feature weak target detection algorithm as the baseline,the analysis and comparison are done.Finally,based on the multi-fractal characteristics of the sea echo,it is proved that the Auto Regressive(AR)power spectrum of the sea echo also has multi-fractal characteristics,and a weak target detection algorithm based on the echo of mean square summation local multi-fractal spectrum of AR power spectrum is proposed.Combined with Support Vector Machine(SVM),a weak detection algorithm based on SVM and multi-fractal feature is proposed.By comparing detection algorithms,it is proved that the detection performance of this algorithm is the best.