Research On Detection Of Targets In Sea Clutter

Sea target detection is an important part of large-scale maritime monitoring,detection and early warning,and it is also a hot spot and key issue in radar signal processing.Due to the complex characteristics of sea clutter,it is particularly difficult to detect floating,low-speed small targets in its background.This thesis excavates the characteristics of sea targets and sea clutter on the basis of the sea surface echo data measured by sea detection radar.Research on target range unit detection and target signal extraction with low signalto-clutter ratio.The specific work and innovations are as follows:1.The time correlation and amplitude statistical characteristics of sea clutter are studied based on the two kinds of measured data sets of IPIX measured data and Seadetecting X-band Radar and Data Acquisition Program,and the basic model of target signal echo on the same distance unit is constructed.2.Aiming at the problem of target range unit detection under single polarization data,a target range unit detection method based on sequence features is proposed from the perspective of feature extraction and classification.This method extracts sequence features in different feature domains,and combines Bi-directional LSTM with memory characteristics to train and detect them.The false alarm probability is controlled by a custom threshold,and the target distance unit detection with controllable false alarm is realized.The verification of the measured data shows that the method effectively improves the detection probability of the target range unit and shortens the detection time.3.Aiming at the problem of multi-polarization data joint target distance unit detection,a target distance unit detection method based on joint polarization sequence features is proposed.This method combines the weighted fusion of features of each polarization data with bidirectional LSTM to achieve target distance unit detection.The measured data show that the joint detection performance of HV and VH dominant is the best,and it is better than the single polarization detection performance.4.Aiming at the problem that it is difficult to detect the slow sea target signal in the low signal-to-clutter ratio radar echo,based on the fractional Fourier transform,combined with the ensemble empirical mode decomposition,a sea surface slow and weak target detection method based on EEMD-FRFT is proposed.This method uses the decomposition characteristics of EEMD to separate a part of sea clutter by dividing and discarding highfrequency IMF components,and then combines FRFT to realize the detection of slow and weak targets.Using the measured data and simulation targets,it is proved that the method can still detect the slow target signal at low signal-to-clutter ratio,and can accurately estimate the center frequency and frequency modulation of the target signal.In summary,this thesis studies the problem of target distance unit detection under complex sea conditions and slow and weak sea target detection.The target distance unit detection based on sequence features and the slow and weak target detection method based on EEMD-FRFT are proposed respectively.The performance of sea target detection in complex marine environment is effectively improved,and it has good application value and prospect in accurately detecting and monitoring sea surface targets.