Research On Compressive Sensing Method For Maritime Target Detection

Due to the impact of sea clutter,conventional maritime target detection method is easy to produce high false alarm problems.How to suppress non-stationary,strong cor-relation sea clutter effectively,and improve the detection capability,has been a difficult problem in the field of radar detection.In the case of high frequency approximation,the backscattering of sea targets(such as ships,etc.)often exhibits the characteristics of multi-scattering center.The number of scattering centers is generally much smaller than the number of distinguishable units in the observation area,which basically meet the compressive sensing(CS)approach to the target scattering sparse a priori require-ments.Therefore,this paper focuses on the use of compressive sensing technology to carry out the following research work:1.In the background of the Gaussian noise,we studied the object refactoring capa-bility and output noise characteristics of the Iterative Soft Thresholding algorithm.The two types of fixed threshold detectors based on IST are discussed,and the analytic ex-pressions of the probabilities and false alarm probabilities are deduced.The simulation results show that the IST fixed threshold detector under the under sampling conditions is better than the matched filter optimal detector.In addition,the architecture of CFAR detector-based on IST and its performance analysis are given.2?Under the condition of sea clutter background and point target detection,the sea clutter composite K distribution model was established.Then,we investigated the sup-pression performance of Orthogonal Matching Pursuit(OMP)and Focal Undetermined System Solver(FOCUSS)algorithms for low sea situation.Compared to the classic whitening filtering method,they have better sea clutter filtering ability.Aiming at the complex dynamic behavior of sea clutter,the deep learning method of sea clutter sup-pression is also explored,we used the convolutional self-encoder(CAE)to effectively isolate the sea clutter and the target in the echo spectrum,and verify the feasibility of the method.3?We did research on extended target detection in the background of the sea clutter.The multi-scattering centers of the extended targets usually exhibit the characteristics of continuous regional distribution.In this paper,the SF-LASSO algorithm with the regularization constraint of Total Variation(TV)is proposed by using the sparseness of the boundary of the target continuous region and the continuous dependency between the non-zero target scattering point and the surrounding scattering point.The simulation results show that the SF-LASSO algorithm can accurately invert the target position and the basic contour.