Feature Extraction And Recognition Of Middle Ballistic Target Group In Complex Environment

As the form of the modern information battlefield is becoming more and more complex,the way to identify and intercept ballistic missiles through information technology has become extremely important.Various countries have rushed to develop various advanced feature extraction and identification methods.In the mid-flight of the ballistic missile,in addition to releasing the warhead,it also releases the target group,which mainly includes light and heavy decoys.The main task of anti-missile work is to accurately identify real warheads and fake decoys to lay the foundation for rapid and accurate interception.The main work of this paper is as follows:In the complex scene of the middle section,research and establish the motion estimation modeling and solution formula of the middle section of the trajectory,introduce the common fretting forms of the space target,and analyze the fretting under the conditions of the general scattering center and the equivalent scattering center of the space target.Equation,simulation realizes the broadband echo and original echo of the mid-range target groupThe problems of Doppler frequency estimation and echo compensation of the ballistic projectile in the middle of the trajectory are studied.Based on the comparison,the micro-motion feature extraction method based on multi-target tracking(MTT)is first used.Use echo phase information to make time-frequency changes,and use MTT technology to track the time-frequency distribution curve.In order to make the frequency estimation not affected by the window length,in order to obtain the Doppler frequency of the target,the estimation method under the time-frequency autoregressive model is studied,and the power spectral density function after the echo parameterization is obtained,based on the Kalman under the CWPA model Under filtering,the error frequency value in the estimation process is eliminated and the re-association in the case of error association is analyzed at the same time.After the frequency point estimation of the scattering point is completed,the Doppler defuzzification processing is performed.The defuzzification process is mainly completed by the way of spectrum shifting.Using the obtained envelope Doppler curve,the polynomial fitting loop compensation fitting is used to obtain The final jog frequency curve after target compensation.Using the obtained equivalent scattering center lower middle target group as an example,the feature extraction and classification of the middle target group are studied.First,the Doppler area is divided by the energy intensity of the time-frequency distribution,combined with the time-frequency distribution differences of different targets,four kinds of micro-motion Doppler-related features are extracted to lay the foundation for classification.And it lays the foundation for the subsequent calculation of the precession parameters of the cone warhead using the HRRP sequence.The space cone target contains a lot of information and parameters when it completes the precession micro-motion in space.Three solutions are mainly analyzed to solve the parameter estimation problem of the cone warhead:Single-view research requires auxiliary trajectory prediction information.The cone spin frequency is extracted using the cyclic correlation algorithm.On this basis,the super-resolution algorithm is used to extract the macro changes of the scattering points,and the cone is derived with the aid of the trajectory prediction technology.The parameter estimation expression of the warhead.Under the condition of echo coherence,the extraction of target velocity and acceleration information is further completed.When such auxiliary trajectory information cannot be obtained,parameter estimation under the background of dual perspectives comes into being.Under different observation time ranges,multiple sets of micro-motion distance sequence changes are obtained,and multiple sets of equations are constructed,so that the relevant parameter estimation can be completed when the trajectory information is unknown.In the actual system,the signal-to-noise ratio will have an impact on the macro extraction of the curve.If the signal-to-noise ratio is low and cannot be used to clearly extract the macro changes of the scattering points,the parameter derivation of the single-view or double-view of the above algorithm will Invalidate.In order to solve the situation that the signal-to-noise ratio is low and the echo is incoherent and the scattering point macro changes are difficult to extract,we introduce an improved generalized Radon transform for parameter estimation and extraction.Due to the existence of this effect,the number of scattering points in different viewing angles is different under different observation situations.We have classified and discussed the derivation and estimation formulas under different situations,and pointed out that the fracture curve can also be used for effective parameter estimation.