| Synthetic Aperture Radar(SAR)achieves two-dimensional coherent integration by transmitting wideband signals and exploiting synthetic aperture,which is able to obtain highresolution images of remote ground scene with all-day and all-weather capability,and is widely used in military and civilian fields such as battlefield reconnaissance,resource exploration,environmental monitoring,etc.However,with more radio equipment,the electromagnetic environment is becoming more and more crowded,and the SAR system is inevitably interfered by radio frequency interferences from equipment in the same frequency band in practice.Among them,the frequency band of wideband radio frequency interference(hereinafter referred to as “wideband interference”)overlaps a lot with the operating frequency band of the SAR system.The existence of a large number of wideband interferences will lead to bright patches and stripes in the SAR image,and even complete coverage of the target,seriously degrading the quality of the SAR image,further affecting the subsequent target detection and recognition,and greatly reducing the SAR system’s ability to obtain scene information.Because wideband interference overlaps a lot with target echo in time and frequency domain,and exhibits time-variant and non-stationary characteristics,how to effectively detect and accurately remove wideband interference is a hot and difficult topic in the current SAR signal processing field.Therefore,the research on SAR wideband interference suppression methods will effectively improve the utilization of SAR echo data and the information acquisition ability of SAR system in complex electromagnetic environment,which has extremely important theoretical value and broad application prospects.Based on the research of non-stationary signal processing methods,combined with the wideband interference received by the practical SAR system,this dissertation focuses on investigating the accurate description,effective detection and accurate removal of wideband interference,and obtains corresponding research results.The main contents are listed as follows:1.As the research basis of interference suppression methods,this dissertation establishes the SAR echo signal model in the presence of wideband interference firstly,analyzes the characteristics of wideband interference in the time domain,frequency domain,and timefrequency domain,and theoretically reveals the impact of wideband interference on imaging results of the SAR system.Then,from a statistical point of view,the differences between wideband interference and target echo are analyzed in time-frequency domain,and then a wideband interference detection method based on adaptive threshold is proposed.The method adopts the hypothesis test of a distribution to determine the non-interference area on the time-frequency spectrum,and then its probability distribution is utilized to calculate a reasonable detection threshold,which effectively solve the problem that it is difficult to find the interference-free echo to calculate the detection threshold in the case of a lot of interferences.In addition,aiming at the problem that the echoes of strong scattering points may affect wideband interference detection and suppression,an echo protection method for strong scattering points based on the correlation of adjacent echoes is proposed.When there are some strong scattering points in the scene,the echoes of strong scattering points are extracted from the original echoes first,so as to avoid the problem of removing the echoes of strong scattering points by mistake as wideband interference when performing interference suppression,i.e.,protecting the echoes of strong scattering points.The experimental results verify the effectiveness of the proposed wideband interference detection method and the echo protection method for strong scattering points.2.Aiming at the problem that single-component high-power wideband interference is difficult to suppress due to its time-variant and non-stationary characteristics,a wideband interference suppression method based on instantaneous phase estimation,demodulation,and filtering and a wideband interference suppression method based on variational complex mode decomposition are proposed.First of all,the relationship between the instantaneous phase of wideband interference and the phase of its time-frequency spectrum is derived,and then the instantaneous phase estimation and demodulation method based on time-frequency spectrum is proposed to reveal the phase change law of wideband interference,and its energy is highly concentrated near the zero frequency for the convenience of subsequent filtering.Next,in order to accurately determine the endpoints of the interference pulse,an interference endpoints estimation method based on correlation coefficient is proposed to effectively determine the location of the interference in time,and then the interference is removed by filtering in frequency domain.In addition,the nonlinear chirp mode decomposition method is extended to the complex field,and the wideband interference suppression method based on the variational complex mode decomposition is proposed by taking the wideband interference as the signal to be extracted.This method establishes a variational framework,iteratively estimates the instantaneous frequency of wideband interference and adaptively extracts the interference components by minimizing the bandwidth of interference after demodulation,so as to achieve the purpose of interference suppression.Finally,the experimental results of simulated and real measured data show the effectiveness of the proposed instantaneous phase estimation,demodulation and filtering method and the variational complex mode decomposition method in single-component wideband interference suppression.3.Faced with the challenge that multi-component wideband interference is difficult to suppress due to its complex and diverse forms and even overlapping in time-frequency domain,the window width optimization method for the short-time Fourier transform(STFT)based on concentration measure,the instantaneous frequency estimation method based on Viterbi algorithm and the regularized time-frequency filtering model are proposed.First of all,the window width is searched efficiently in the logarithmic scale to achieve the optimal value of the concentration measure of the time-frequency spectrum,and the highly concentrated time-frequency spectrum is generated to provide a good basis for subsequent processing.Next,Viterbi algorithm is used to detect the ridge path of time-frequency spectrum,and re-detect the ridge path according to the preliminary obtained ridge path to avoid the impact of the previous component detection process on the subsequent component detection process.If the ridge paths intersect,they are reconnected at the intersection point according to the similarity of the slope of the ridge paths on both sides,and the final ridge paths are taken as the estimated values of the instantaneous frequencies.Then,a regularized time-frequency filtering model based on instantaneous frequency is established.This model not only considers the accuracy of interference reconstruction,but also takes advantage of the sparsity of wideband interference in time-frequency domain,so as to minimize the loss of useful target echo while suppressing wideband interference.Finally,the experimental results of simulated and real measured data verify the effectiveness of the proposed method.4.Aiming at the problem that multi-component wideband interference is difficult to remove accurately due to the rapid change of instantaneous frequency and even serious overlap in time-frequency domain,a wideband interference suppression method based on masked second-order multisynchrosqueezing transform is proposed by combining second-order synchrosqueezing transform and multisynchrosqueezing transform.Firstly,the initial instantaneous frequency estimate and the instantaneous chirp rate estimate are obtained from the time-frequency spectrum,and the window width of STFT is optimized to make the timefrequency spectrum more concentrated,and the time-frequency mask corresponding to the wideband interference is obtained by interference detection.Next,the second-order instantaneous frequency estimate is calculated based on the instantaneous frequency estimate and the instantaneous chirp rate estimate obtained from the optimized timefrequency spectrum,and the points of the time-frequency spectrum are reassigned to concentrate the energy of wideband interference on the instantaneous frequency curve.This process can be repeated several times to make the result more concentrated.This step is referred to as the second-order multisynchrosqueezing transform.Then,the time-frequency mask of wideband interference is refined according to the transform result,and it is multiplied with the transform result in an element-wise manner to obtain the masked transform result,and then the wideband interference extraction result in the time domain is obtained by inverse transform.The echo after interference suppression can be obtained by subtracting the extracted wideband interference from the original echo.Finally,the experimental results of simulated and real measured data show that when the instantaneous frequencies of wideband interferences change rapidly,even overlap in the time-frequency domain seriously,this method can still accurately suppress interference and retain the energy of useful target echoes well. |
PDF Full Text Request