Research On Feature Extraction And Quality Evaluation Of Radar Imaging For Space Target

Inverse Synthetic Aperture Radar(ISAR)is a kind of radar with high resolution imaging capability which is different from traditional radar.ISAR is usually a stationary radar that performs two-dimensional high-resolution imaging of moving targets vertically and horizontally,so it can obtain important information about the size,shape,structure of spatial targets such as satellites and space stations.It plays an important role in national defense security,acquisition of war information and target attack.With the development of military technology,the research on feature extraction and quality evaluation of ISAR imaging has important military value.In the aspect of feature extraction,the battlefield environment is more and more complex,the maneuverability and disguise ability of military targets are increasing,and it is more and more difficult for ISAR to capture target information clearly.Moreover,a single ISAR imaging can only reflect part of the details and characteristics of the target,and it is difficult to get a comprehensive and accurate description of the target.Military image fusion technology can fuse the images of a series of military targets acquired by ISAR to obtain more comprehensive information of the target,and has been widely used in space target recognition,attitude estimation and so on.It is of great military significance to derive the three-dimensional structure of a target from its two-dimensional information.Accurate threedimensional reconstruction of target can make target indication more accurate and play an important role in military surveying and mapping,unmanned combat,military rescue,etc.In the aspect of quality evaluation,image quality is an important index to measure the overall performance of radar system.The quality evaluation of ISAR image is very important for performance evaluation of radar system,analysis and comparison of algorithms,etc.It can be used for relevant researchers to detect and design radar performance,and technicians to improve the algorithm.The real-time and accuracy of feature extraction and quality assessment of ISAR imaging in practical applications are the hot spots of research.In this paper,two aspects of information processing are studied: two-dimensional multi-view fusion of high-resolution ISAR images and three-dimensional reconstruction of targets.At the same time,quantitative evaluation methods of image quality are studied for high-resolution ISAR images.The main contents of each section are as follows:The first part mainly solves the two-dimensional and multi-view fusion of high-resolution ISAR images.Using the structure information of the image,a series of ISAR images are segmented by Simple Linear Iterative Clustering(SLIC),and a rigid rotation matrix between ISAR images is constructed based on the super-pixel center to register and fuse all ISAR images.Experiments on ISAR images of sequential space shuttles verify that this method can accurately extract the effective feature points needed for registration and has a high fusion accuracy.The second part mainly solves the problem of target three-dimensional reconstruction of high-resolution ISAR images.This paper introduces a statistical filtering method to preprocess two-dimensional ISAR images,uses the Kanade-Lucas-Tomasi Tracking(KLT)algorithm to extract and track image feature points,and uses the Orthographic Factorization Method(OFM)to reconstruct a series of ISAR images in three dimensions.Experiments on ISAR images of sequential space shuttles show that this method balances the computational complexity of solving feature points with the accuracy of three-dimensional reconstruction,and is widely applicable and efficient.The third part mainly solves the ISAR batch imaging quality evaluation problem.First,the Support Vector Machine(SVM)classifier is briefly introduced.Then,the method of supervisory feature extraction using Convolutional Neural Networks(CNN)is introduced.Finally,a batch imaging quality evaluation method for ISAR is presented,which combines subjective evaluation with objective evaluation.Experiments on the ISAR simulation image of Tiangong No.1 verify that this method is practical and has high accuracy.