Research On High Frequency Band Radar Target Features Extraction And Identification

Radar technology research focus on target recognition recently.Ship and airplane targets locate in the resonance region for the working frequency of high frequency band radar,an important feature of the resonance region is characterized by extreme,this pole is characterized by the inherent properties of the target,it is the most effective characterization of radar target characteristics.Therefore,the study of high-frequency band radar has a strong practical significance target recognition method based on the pole.Target recognition method based on pole focused target acquisition and analysis of two studies on the pole of recognition.Therefore,this article will get radar target scattering characteristics,high-frequency radar target feature extraction,high frequency radar target recognition studying four aspects.Firstly,it relly mainly on pole acquisition method simulation,which led to the extraction accuracy of the pole and simulation methods are closely related,so a precise modeling and calculated methods are needed.Guidelines conventional modeling methods do not have a determined operator-dependent modeling of subjective consciousness,which can cause errors.This is the first high-frequency radar modeling guidelines given complex goals,and put forward the fast multipole method and the method of Mo Ments combined solution of equations of the new algorithm FMM-Mo M,with calculation software Feko parameter settings to achieve the software directly.By using this method,effectively improve the accuracy and computation time.Secondly,to obtain a precise target scattering characteristics,the next step is to study how to get its pole features.Methods have been early and late response aliasing problems and late response to the start time is determined empirically set,will cause extreme extraction inaccurate.This article will use ARMA model algorithm to extract features to avoid extreme early and late response aliasing,because the previous parameter settings are based on experience and can only give an approximate range,too subjective cause extreme extraction inaccurate.To improve the previous ARMA algorithm itself because of defects and extreme dependence on the parameters extracted caused too inaccurate shortcomings.Finally,got a new and improved ARMA model algorithm.For reliable extraction of target aircraft and ships pole features.In fact only one direction.Targets also extracted false pole and missed the real possibility of objective truth can not guarantee that all of the poles will be able to extract.Therefore it is necessary to integrate azimuth pole,the pole may finally get a group of representatives of all directions, achieving pole orientation has nothing to do with the target.This paper presents an integrated approach using feature extraction pole and multi-directional correction method combined to obtain a more accurate new extreme.Followed by complex target ships and aircraft extreme feature extraction,and the first analysis of the impact of changes in the length and breadth and complexity of the various parts of the objectives of the poles.Then,frequency selected method was studied based on residue to identify new targets.To keep the number of properties on the basis of the pole based on a theoretical analysis,for the first time by the number of changes in the value of the simulation of the impact of leaving the pole to verify the number of really stay with the polarization angle dependent,but to get the number and size of the left pole verified proportional to the size,and function of the number of poles corresponding to remain associated with the polarization angle,and polarizationrelated,leaving a few very small or very close to zero,as well as its corresponding pole unable to be extracted.Based on high-frequency radar target multi-frequency characteristics of the target classification and recognition method,the key to this method is to find the right frequency,and therefore proposed to remain the most preferred method is based on the maximum number of frequency method R-SVM.By identifying aircraft and ship targets simulation verified that the method can achieve effective recognition of the target database.At last,it processed how to use this target recognition method based on poles in engineering practices.First,the impact of the ship rocking and sea state to carry out simulation using simulation modeling simulation of the rotation angle of the ship swing,move up and down and left and right tilt to simulate the waves,the simulation data obtained were poles feature extraction,fusion in the database target classification and identification,and target a large number of aircraft and ships to identify random simulation is used to validate the correctness of the theory and methods R-SVM-TNN.Achieved a target ship for the library on the basis of reliable identification of non-library judge refused to ship targets effectively.