Noise Suppression In Sea Clutter And Its Weak Signal Detection

Noise suppression and weak signal detection in sea clutter play an important role in ocean monitoring,national economy and security.In the complex Marine environment,weak signals are often drowned in strong sea clutter.Traditional detection methods are prone to interference from environmental noise and ignore the internal movement law of sea clutter,resulting in weak characterization ability and low detection accuracy.It is of great theoretical significance and application value to improve the sea surface monitoring ability of sea detection radar to study the characteristics of sea clutter and put forward a new radar echo modeling method to improve the detection technology of weak signals to identify floating small targets.In this paper,the characteristics of weak signals in the background of sea clutter is studied,noise suppression methods in sea clutter is proposed,the HFER feature classification network of weak signals is designed,and the image representation of the correlation difference between clutter and signal is analyzed,deep learning is introduced into weak signal detection,the sea clutter two-dimensional detection model and chaos prediction model is established to provide a theoretical method for the effective detection of weak signals in the sea clutter background.The specific research contents are as follows:(1)Aiming at the interference of noise to weak signal detection in sea clutter,MTCEEMD noise suppression method combining CEEMD and wavelet packet multi-threshold processing is proposed.According to the autocorrelation function and energy ratio of IMF decomposed by CEEMD,high-frequency IMF is classified.The wavelet packet decomposition coefficients of high-frequency IMF are arranged in frequency order,and the appropriate threshold filtering is selected according to the characteristics of frequency band.The sea clutter after denoising is reconstructed with the unproposed low-frequency IMF.Sea clutter prediction model is established by Volterra adaptive filter and comparative experiments are carried out.(2)In order to improve the performance of weak signal detection by sea radar,a weak signal detection method based on HFER feature is proposed.The target distribution characteristics of high frequency IMF energy ratio are studied,which are regarded as features to design weak signal detection network based on XGBoost,and the network hyperparameter set is optimized by SSA.By comparing the detection probability of BCD,RAA,TIE,SVM and multi-feature fusion detection methods,it is verified that HFER feature is suitable for weak signal detection in sea clutter background.(3)In order to solve the problem of difficult representation of one-dimensional sea clutter information,a two dimensional detection method of sea clutter is designed.The image representation method of the respective correlation between sea clutter and weak signal is studied,and the binary classification of clutter and signal is carried out by Alex Net network.Three image conversion channels are compared,data cleaning is carried out through MTCEEMD and the influence of variable step size parameters on network performance is analyzed.The 10^-3false alarm controllable premise is set and compared with the figure with density method,Hurst index algorithm and time-frequency three-feature detection method.(4)In order to accurately locate the existence range of weak signals in sea clutter,a chaotic sea clutter prediction model based on PSR-LSTM is established.The chaotic characteristics of sea clutter is studied,the phase space reconstruction theory is combined with LSTM network,the width of embedded window is used to determine the network training step size,so that each group of training data and sea clutter keep the same topological structure,and a detection method of prediction error frequency domain transformation is proposed.RBF neural network,SVM,LS-SVM,GA-SVM and ELM methods are compared to carry out validation experiments.The motion law of sea clutter is analyzed,denoising pretreatment layer is designed,feature extraction method is proposed,weak signal detection model and sea clutter prediction model are established.It solves the problems that the traditional method is susceptible to a variety of interference and noise in the complex Marine environment,and the characteristics of weak signals are not accurately matched,and the detection ability is weak.It provides a strong support for the improvement and development of sea radar detection technology,and has good theoretical significance and practical value...