The Study On Scattering-Mine Detection In High-resolution SAR

With the accelerating pace of modern high-tech local War battlefield,scattering-mine will be used more and more widely,but the mines left over from the war threaten the people lives and limit the economic development,therefore,for the reason of the need of war and humanitarianism,the research on scattering-mine detection technology must be developed rapidly.Meanwhile,with the rapid development of mini SAR technology,the advantage of mini SAR in mine detection field gradually emerges,developing mine detection method based on mini SAR platform becomes a trend.As the key link,the technology of throwing mine detection in high-band and high-resolution SAR has become an important research direction.Scattering-mine detection is a typical problem of small target detection in heavy cluster,in order to achieve the effective detection,the problem of clutter distribution model selection,target extraction and detection algorithm design must be solved.Therefore,focusing on these three aspects of scattering-mine detection,exploration and research are carried out in this paper.Specific contents and results of research are as follows:(1)Firstly some clutter distribution models and parameter estimation methods are introduced and analyzed,fitting experiments on several typical nature scene clutter of the measured data are did and experimental results are assessed then,furthermore,the modeling capability of several clutter distribution model to different scenarios are analyzed.(2)The target azimuth scattering characteristics can be obtained by sub-aperture SAR images,therefore,the generation method of SAR sub-aperture image is introduced,and the problem of uneven distribution of echo energy among the sub-apertures is mainly solved;Based on measured data,extract the target amplitude and phase information from the sub-aperture image sequence to analyze the azimuth invariance and the azimuth coherence of the target.The result shows that the azimuth coherence feature can be used to clearly distinguish target and clutter.(3)CFAR detection algorithm based on clutter distribution model has been deeply studied and widely used,and the principle and detection process of it are introduced in this paper;Sub-aperture coherence detection algorithm is based on the complex image data,which mainly employs the coherence of targets and clutter.It is found from experiment that detection algorithm based on single feature is not robust.Combination with CFAR detection and sub-aperture coherence detection,the information fusion detection method based on local contrast and azimuth coherence feature is introduced and analyzed.Experimental results show that this method can improve the detection capability,and it also verifies the feasibility and validity of azimuth coherence in high-band and high-resolution SAR scattering-mine target detection.