Multi-frame Detection Method Of Small Targets In Sea Clutter

With the development of electronic information technology,radar detection technique of sea surface targets has become more and more mature.Existing detection methods have attained rather good performance to traditional warships and fishing boats.However,the detection of sea-surface small targets remains one of difficult problems in radar signal processing field.Sea-surface small targets usually have the following characteristics.1)small radar cross-section area,such as periscope and buoy.2)irregular target appearance time on sea surface,for instance,periscope,air vent,communication buoy of a submarine and other small targets,may only last for several seconds.3)unstationary Doppler characteristics.Slow motion of small targets and their complex interaction with sea waves complicates their Doppler offset change with time.Moreover,in high-resolution case,there exist a large number of sea spikes similar to small targets in characteristics,which are difficult to distinguish sea spikes and small targets in a short observation time.These factors make it a long-term problem to detect sea-surface small targets effectively in sea clutter.In this thesis,we devote to investigation of effective detection methods and algorithms of sea-surface small targets by exploiting substantial differences of sea clutter,sea spikes,and small targets in multiple scans in high-resolution and fast scan operating mode of radar.The main content of this thesis is summarized as follows: In the second chapter,three kinds of traditional and widely-used sea surface target detection methods are reviewed,including incoherent CFAR detection method,optimum coherent detection method based on the generalized Pareto sea clutter model and feature-based detection method.Incoherent CFAR processing is simple but its detection ability is limited for sea-surface small targets.The optimum coherent detection method based on sea clutter statistical model exploits refined signal processing and thus attains good detection performance sea-surface moving targets.However,it is difficult to distinguish between sea spikes and small target echoes because of limited integration time.The feature-based detection method is an effective way to find sea-surface small targets but its application is limited due to long integration time.In the third chapter,the characteristics of sea clutter,sea spikes and small targets are analyzed based on raw radar data at high resolution fast scanning mode.The analysis shows that high-resolution sea clutter data follows the generalized Pareto clutter model.In addition,temporal and spatial correlations of sea clutter and characteristics of sea spikes are discussed.Based upon these analysis,a scheme of multi scan joint detection is given for sea-surface small target detection in high-resolution fast scan mode.The fourth chapter focuses upon the detailed flowchart of the detection method and experiments using simulated and measured data.The detection method is realized in two cascaded steps.At the first step the processing in a scan uses the traditional detection method to find out targets and part of sea spikes using a lower threshold.Its aim is to temporarily ignore the false alarm from sea spikes while ensuring effective detection of small targets of interest.At the second step,the multi scan processing exploits multi scan integration based upon the retrospective filter bank to exclude false alarms from sea spikes that fail to excluded in the first step.The retrospective filter bank exploits the difference of inter scan motion characteristics between sea spike and small targets to effectively integrate target traces while preventing integration of the false alarms from sea spikes.At last,the proposed detection method is verified by using simulated and real sea clutter data.The fifth chapter is the summary of this thesis and a prospect of future research.