Radar Target Detection Method And Implementation Based On Ground/Sea Clutter Scene Identification

Target detection under sea clutter conditions plays a crucial role in both military and civilian fields.However,this research area faces two challenges.First,the traditional CFAR detector is mismatched with the statistical model of sea clutter,which leads to poor performance of the detector.Second,the presence of multiple strong clutter or interference targets adjacent to the actual target can cause the detection threshold to be too high,resulting in target masking and reduced performance of the traditional CFAR detector.To address these challenges,this paper conducts in-depth research on the distribution model of sea clutter and the relevant theory of CFAR detection,starting from the background characteristics of land/sea clutter and the design of the detector.Based on this,a series of methods to improve the performance of target detection at sea are proposed,and the radar target detection process is implemented through a GPU development kit.Therefore,the specific research content of this paper is as follows:(1)In response to the radar sea detection work scenario,this paper deeply analyzes several typical sea clutter amplitude distribution models and briefly introduces the measured data used in this paper.Secondly,to solve the problem of performance limitation of traditional CFAR detectors due to the mismatch between the statistical model of sea clutter and the detector,a sea radar target detection method based on scene recognition of land/sea clutter is proposed.This method identifies the radar working environment and recognizes the sea surface scene as land island area,nearshore clutter area,clutter noise mixed area,and far distance noise area.Finally,the sea clutter in each region is fitted through kernel estimation method to provide theoretical support for the selection of suitable detectors in the subsequent detection process.(2)To address the limitations of CFAR detectors in the detection process,an improved CFAR method utilizing phase characteristics to screen reference cells is proposed.This method determines the phase linearity of consecutive pulse echoes in the same reference cell and removes reference cells with strong phase linearity,forming a detection threshold with better adaptability to the target cells.The algorithm is compared with other mean-based CFAR algorithms in different environmental backgrounds through simulations and measured data tests.Results show that in multi-target and clutter edge environments,the improved CFAR algorithm can better avoid target masking and has better detection performance than other mean-based CFAR algorithms.(3)In view of the engineering application prospects of the research content,a GPU development kit was selected to implement the complete process of radar target detection from data reading to target detection,including clutter suppression module,non-coherent accumulation module,CFAR detection module,and target tracking module.The GPU development kit displays the processed results of the radar scan data in modules and compares them with the results processed in Matlab to demonstrate their accuracy.At the same time,the acceleration ratio is reflected by comparing the CPU running time.