Research On Efficient Target Detection Method In Complex Scene Of Airborne SAR

With the rapid development of synthetic aperture radar(SAR)technology,the efficient acquisition and interpretation of airborne SAR image information has become a research hotspot in recent years.As the key research content of airborne SAR image interpretation,target detection is the key link to accurately locate the target area of interest,which has important application value and research significance.In the actual environment,large scene SAR images contain rich ground object information,and the scene is complex,which brings many difficulties and challenges to the detection of interested targets.Aiming at the application requirements of efficient and accurate interpretation of airborne SAR images,focusing on the problem of target detection in complex scene SAR images,the thesis studies the accurate detection of large scene SAR targets and the efficient imaging detection of complex scene SAR.The main research contents are as follows:(1)The basic theory of SAR target detection is studied.Firstly,the characteristics of SAR image are analyzed.On this basis,the basic principle of constant false alarm rate(CFAR)detector in traditional target detection methods is studied.Then,the deep learning-based target detection method is studied,and the two-stage and single-stage deep learning detection algorithms are summarized to provide support for subsequent research.(2)A SAR target detection method in complex large scene is studied.A "slice detection merge" target detection model in complex large scene is established.Based on the residual structure target detection network,combined with the feature fusion module,the multi-layer features of SAR targets are extracted,and the non maximum suppression(NMS)processing is used to remove the redundant detection results in the scene,so as to realize the accurate detection of SAR targets in complex large scene.(3)An efficient imaging detection method for SAR complex scene is proposed.First,the integrated framework of imaging and detection is established.Then,the potential target area in the imaging process is extracted by using fast imaging and significance iterative screening.Combined with the target detection network,the target highresolution imaging and detection results are obtained,and the efficient imaging and detection of SAR target is realized.The proposed methods in the thesis have been validated by simulation experiments.The experimental results show that these methods can effectively solve the problems in SAR target detection under the above complex scenarios,and provide methods and technical support for airborne SAR target detection.