CN-122017749-A - Priori time-frequency information-assisted satellite-borne synthetic aperture radar broadband interference suppression method, system and storage medium

CN122017749ACN 122017749 ACN122017749 ACN 122017749ACN-122017749-A

Abstract

The invention discloses a priori time-frequency information-assisted broadband interference suppression, a system and a storage medium of a satellite-borne synthetic aperture radar, which comprise the steps of selecting reference pulses and constructing time-frequency domain local amplitude characteristics through short-time Fourier transform; the method comprises the steps of extracting the frequency band and the duration of interference by an absolute medium bit difference method based on reference pulses as prior information, dividing a time-frequency sub-matrix containing the interference pulses, distinguishing strong scattering areas and weak scattering areas, respectively carrying out K-means clustering to primarily distinguish the interference, carrying out constraint and correction on a primary result by combining the prior information, improving the integrity and consistency of interference positioning under a complex background, and finally adopting block subspace filtering to inhibit interference components in the interference areas. The invention adapts to different scenes by introducing the interference priori and time-frequency division control mechanism, effectively inhibits the broadband radio frequency interference, reduces the damage to useful signals and is beneficial to improving the imaging quality.

Inventors

LI NING
LI ZIJUN
HUANG YABO
WU LIN
ZHAO JIANHUI
SHU GAOFENG

Assignees

河南大学

Dates

Publication Date: 20260512
Application Date: 20260130

Claims (10)

1. The prior time-frequency information-assisted satellite-borne synthetic aperture radar broadband interference suppression method is characterized by comprising the following steps of: Step S101, performing pulse-by-pulse distance Fourier transform on data containing interference echo, calculating spectrum flatness, sorting according to the size of the spectrum flatness, and selecting a plurality of pulses with the lowest spectrum flatness as reference pulses; step S102, short-time Fourier transform is carried out on the reference pulse based on the reference pulse, a plurality of submatrices are generated in a time-frequency domain, and local amplitude characteristics of the time-frequency domain are constructed; Step S103, positioning an interference area by an absolute medium bit difference method based on the time-frequency local amplitude characteristics of the pulse, extracting an occupied frequency band of interference in a time-frequency domain and the duration of the interference, and constructing interference priori information; step S104, dividing a time-frequency representation containing interference pulses into a molecular matrix and calculating local amplitude characteristics based on the method in the step S102, and dividing the time-frequency representation of the pulses into a strong scattering region and a weak scattering region by adopting an absolute medium bit difference method; step S105, respectively carrying out K-means clustering on the local amplitude characteristics of the time-frequency domain in the strong scattering region and the weak scattering region to realize preliminary distinction between an interference region and a non-interference region, and obtaining an initial positioning result of time-frequency domain interference; Step S106, combining the interference priori information constructed in the step S103 to restrain and correct the initial interference positioning result, and supplementing and correcting a time-frequency communication structure meeting the interference priori condition at the junction position of the strong scattering region and the weak scattering region to obtain a final time-frequency domain interference positioning result; And step S107, performing interference suppression processing on the finally determined time-frequency domain interference positioning area by using a block subspace filtering method, and outputting echo data after interference suppression as a final interference suppression result.
2. The prior time-frequency information-assisted satellite-borne synthetic aperture radar broadband interference suppression method according to claim 1, wherein in step S101, pulse-by-pulse distance fourier transform is performed and spectrum flatness is calculated, and a plurality of pulses with the lowest spectrum flatness are selected as reference pulses to extract broadband interference components with prominent interference characteristics, specifically: To measure the uniformity of the echo spectrum energy distribution, let And Representing the number of sampling points in the distance direction and the azimuth direction respectively, and carrying out distance Fourier transform pulse by pulse to obtain the frequency domain representation of the frequency domain representation: ; Wherein, the Represent the first A sequence of time-domain samples of the bar pulse in the distance direction, Indexing to the sampling point for the distance; representing a discrete fourier transform operator; for the corresponding frequency-domain complex spectral components, Represents the frequency index and calculates its magnitude spectrum: ; Wherein, the Represent the first Pulse at the first The spectral flatness SPECTRAL FLATNESS, SF for short, is calculated from the amplitude values at the frequency points, which is defined as the ratio of the geometric mean of the amplitude spectrum to the arithmetic mean: ; To correspond the interference pulse-containing frequency spectrum flatness Sorting, selecting the smallest front The bar pulses act as a reference pulse set.
3. The method for suppressing wideband interference of a priori time-frequency information assisted spaceborne synthetic aperture radar according to claim 1, wherein the step S102 specifically comprises the steps of setting the reference pulse set obtained in the step S101 as: wherein Represent the first The strip reference pulse echoes the signal in the time domain in the range direction, Performing time-frequency analysis on the reference pulses one by one, and processing the reference pulses by using short-time Fourier transform, wherein the time-frequency representation can be expressed as: ; Wherein, the In order to analyze the window function, And Respectively representing time index and frequency index, and obtaining amplitude value of the described time-frequency representation result to obtain correspondent time-frequency domain amplitude distribution The method is used for representing the distribution condition of signal energy in the time and frequency dimensions; after obtaining the time-frequency domain amplitude distribution, dividing a time-frequency representation result in the time-frequency plane according to a preset time scale and a frequency scale, and dividing the whole time-frequency plane into a plurality of time-frequency sub-matrixes which are not overlapped with each other; each time-frequency sub-matrix corresponds to a local area in the time-frequency plane and is used for describing the local energy characteristics of echo and interference in the area, and a first is set The index range of each time-frequency sub-matrix in the time dimension is as follows The index range in the frequency dimension is: then the first is The summation of the amplitudes within the individual time-frequency sub-matrices can be expressed as: ; the statistic is used as a local amplitude characteristic of the time-frequency domain for characterizing the overall level of energy distribution in the time-frequency domain.
4. The method for suppressing broadband interference of a priori time-frequency information aided spaceborne synthetic aperture radar according to claim 1, wherein in step S103, the method is based on local amplitude features corresponding to all time-frequency sub-matrices Dividing the local amplitude characteristics into interference areas and non-interference areas by adopting an absolute median difference method, namely firstly calculating the median of the local amplitude characteristics: ; And on the basis of this, calculating an absolute intermediate potential difference, the expression of which is: ; The segmentation threshold is then determined, set to: ; Wherein the method comprises the steps of The submatrix with the amplitude larger than the set threshold is judged as an interference area and the rest is a non-interference area as an empirical coefficient, and the occupation range of the obtained interference area in the frequency dimension can be expressed as a section based on the obtained interference area: ; Wherein the method comprises the steps of Represent the first The starting frequency position of the interference frequency band in the reference pulse, and the corresponding single interference duration in the time dimension can be expressed as: ; Wherein the method comprises the steps of And Respectively representing the starting time and the ending time of the interference in the time-frequency domain; For the frequency band initial position set And duration set Abnormal value processing is carried out by adopting a median-based abnormal value discrimination mode, namely, any parameter set is subjected to Calculating the absolute intermediate potential difference and eliminating the absolute intermediate potential difference to meet the following conditions: in (2), wherein Is an experience coefficient; After the outlier is removed, carrying out statistics and fusion on the residual effective samples to determine the final interference priori parameters, taking an average value of the starting positions of the interference frequency bands, and carrying out statistics on the duration of single interference in the same way to obtain the final interference priori parameters: ; ; Wherein the method comprises the steps of Indicating the number of effective samples remained after the outlier is removed, and obtaining the frequency band occupied by the interference through the processing Single interference duration Is used for the interference prior information.
5. The method for suppressing broadband interference of a priori time-frequency information aided spaceborne synthetic aperture radar of claim 1, wherein step S104 comprises the steps of, based on the method of step S102, for the first step Pulse at the first The sum of the amplitudes in the time-frequency sub-matrices is as follows Wherein The time-frequency sub-matrix index is represented, For the same echo pulse, the amplitude summation values corresponding to all the time-frequency submatrices form a local amplitude characteristic sequence The local amplitude characteristic sequence is divided into a strong scattering region and a weak scattering region by adopting an absolute medium bit difference method to carry out statistical analysis for distinguishing scattering intensities corresponding to different time-frequency regions, when the amplitude sum corresponding to a certain time-frequency submatrix is larger than a threshold range determined by the absolute medium bit difference, the time-frequency submatrix is judged to be the strong scattering region, and the rest is judged to be the weak scattering region.
6. The method for suppressing broadband interference of a priori time-frequency information aided spaceborne synthetic aperture radar of claim 1, wherein in step S105, step S104 is provided to divide the time-frequency sub-matrix set into a strong scattering sub-matrix set And weak scattering submatrix set For any pulse containing interference echo Taking the corresponding local amplitude sum of the time-frequency submatrices The method comprises the steps of obtaining a time-frequency domain local amplitude characteristic, respectively constructing a sample sequence consisting of the local amplitude characteristic in a strong scattering sub-matrix set and a weak scattering sub-matrix set, automatically classifying the sample sequence by adopting a K-means clustering method in each type of scattering area, setting the clustering class number as 2, and dividing the time-frequency sub-matrix into interference classes and non-interference classes, wherein the objective function of the K-means clustering is defined as the minimization of the square sum of Euclidean distances between the sample and a clustering center to which the sample belongs, namely: ; Wherein the method comprises the steps of Represent the first The number of clusters of the clusters is, Representing the central value of the corresponding cluster in such a way that the mean value of all samples in the cluster is calculated for any local amplitude characteristic sample It is assigned to the cluster corresponding to the cluster center whose euclidean distance is smallest, namely: ; ; after the clustering is completed, the clustering result is judged according to the local amplitude characteristic average value corresponding to each cluster, wherein the cluster with larger local amplitude characteristic average value is judged to be an interference candidate area, and the cluster with smaller local amplitude characteristic average value is judged to be a non-interference area.
7. The method for suppressing wideband interference of a priori time-frequency information assisted spaceborne synthetic aperture radar according to claim 1, wherein the step S106 specifically comprises the steps of: the initial positioning result of the time-frequency domain interference obtained in step S105 can be expressed as a binary indication function Wherein the corresponding time-frequency submatrix is 1 when determined as the interference area, and 0 otherwise, the interference prior information constructed in step S103 comprises the occupied frequency band of the interference in the frequency dimension Duration range of single interference Based on the interference priori information, consistency constraint is carried out on an initial positioning result, namely only a time-frequency region which simultaneously satisfies that the frequency position falls into the interference occupied frequency band and the time continuous length accords with the single interference duration time characteristic is reserved, so that false alarms caused by noise or strong scattering points are eliminated; In the constraint process, when a certain time-frequency area meets the following conditions: ; And confirm the area as the interference area when having the communication relation with the interference area judged in the time-frequency domain, otherwise, for the time-frequency area which does not meet the interference prior condition or only presents the isolated distribution, reject from the interference positioning result; Further, the time-frequency communication structure at the junction of the strong scattering region and the weak scattering region is analyzed, when the time-frequency communication structure meets the interference priori conditions in the aspects of frequency range, time persistence and the like, the corresponding time-frequency region is subjected to supplementary marking, and the communication structure which is obviously deviated from the interference priori characteristics is corrected or deleted.
8. The method for suppressing wideband interference of a priori time-frequency information aided spaceborne synthetic aperture radar of claim 1, wherein said step S107 specifically includes the steps of setting that the interference time-frequency positioning result determined in step S106 corresponds to a block of local data matrix in the pulse time domain representation, and is expressed as: ; Wherein the method comprises the steps of In order for the bandwidth occupied by the interference region, In order for the duration of the interference to be of a duration, The method comprises the steps of constructing a sample covariance matrix for the data block, wherein the sample covariance matrix comprises a useful echo component, an interference component and a noise component: ; and decomposing eigenvalues of the covariance matrix, wherein the eigenvalues comprise: ; Wherein the method comprises the steps of Is a characteristic value matrix and meets , Is a corresponding feature vector matrix; according to the interference positioning result and the characteristic value distribution characteristic, the characteristic vector corresponding to the larger characteristic value is judged as an interference subspace, and the subspace represented by the characteristic vector can be represented as: ; Wherein the method comprises the steps of For the interfering subspace dimension, thereby constructing an orthogonal projection matrix of the interfering subspace: ; Wherein the method comprises the steps of Accordingly, the projection matrix of the non-interfering subspace can be expressed as: Wherein the method comprises the steps of Is a unit matrix; And carrying out filtering processing on the data block by utilizing the non-interference subspace projection matrix to obtain echo data after interference suppression: ; The method comprises the steps of carrying out block subspace filtering processing on the data blocks corresponding to each interference area, projecting and restraining interference components from original data blocks, and simultaneously retaining useful echo components in non-interference subspaces, so that effective restraint on positioning interference areas is realized, and carrying out the processing on the data blocks corresponding to each interference area and reconstructing echo data, thereby obtaining a final interference restraining result.
9. The prior time-frequency information-assisted satellite-borne synthetic aperture radar broadband interference suppression system is characterized by comprising: The interference priori information construction unit is configured to perform Fourier transform on the interference-containing pulse, calculate the spectrum flatness, select a plurality of pulses with the lowest spectrum flatness as reference pulses, then perform short-time Fourier transform on the reference pulses, locate an interference area, extract the frequency occupation range of interference and single interference duration time information, and construct interference priori information; The time-frequency transformation and local dividing unit is configured to perform short-time Fourier transformation on the echo containing interference to be processed to obtain the energy distribution of echo signals in a time-frequency domain, divide a time-frequency representation result in the time-frequency domain to form a plurality of time-frequency sub-matrixes, count the amplitude information of each time-frequency sub-matrix and construct local amplitude characteristics of the time-frequency domain; The time-frequency region dividing unit is configured to analyze the amplitude distribution of each time-frequency sub-matrix by adopting an absolute mid-bit difference method based on the local amplitude characteristics of the time-frequency domain, and divide the time-frequency domain into a strong scattering region and a weak scattering region so as to distinguish echo characteristics under different scattering conditions; The regional interference positioning unit is configured to perform cluster analysis on the local amplitude characteristics of the time-frequency domain in the strong scattering region and the weak scattering region respectively, so as to realize the preliminary segmentation of the interference region and the non-interference region and obtain an initial positioning result of the time-frequency domain interference; The prior constraint correction unit is configured to combine the prior interference information to constrain and correct an initial interference positioning result, and supplements and corrects a time-frequency communication structure meeting prior interference conditions at the junction position of the strong scattering region and the weak scattering region to obtain a final time-frequency domain interference positioning result; an interference suppression unit configured to perform interference suppression processing using a block subspace filtering method for the finally determined time-frequency domain interference location area to suppress broadband interference components and retain useful echo signals; And the pulse merging unit is configured to sequentially merge the echo data subjected to the interference suppression processing, and is used as a final interference suppression result for subsequent imaging processing.
10. A storage device in which a plurality of programs are stored, wherein the program applications are loaded and executed by a processor to implement the a priori time-frequency information assisted method of suppressing broadband interference of a satellite borne synthetic aperture radar according to any of claims 1-7.

Description

Priori time-frequency information-assisted satellite-borne synthetic aperture radar broadband interference suppression method, system and storage medium Technical Field The invention relates to the technical field of signal processing, in particular to a prior time-frequency information-assisted satellite-borne synthetic aperture radar broadband interference suppression method and system. Background Synthetic aperture radar (SYNTHETIC APERTURE RADAR, SAR for short) is an active microwave remote sensing device, which detects ground targets by transmitting microwave signals and receiving reflected waves, generates a virtual large aperture by relying on the motion of a radar platform, and combines a pulse compression technology, thereby realizing high-resolution imaging. SAR signals are not affected by day and night changes and bad weather, and can penetrate through barriers such as cloud layers, rain and fog, so that the SAR signals have all-weather observation capability in all days. With the rapid advancement of modern communication technology, spectrum resources become increasingly stringent. As an open radar system, SAR shares the spectrum of the same frequency band with other devices, which makes the electromagnetic environment in which it is located more complex, and the problem of radio frequency interference (Radio Frequency Interference, RFI for short) is more and more pronounced. RFI can cause disturbance of impulse response, so that problems such as bright lines, fog artifacts and image blurring appear in imaging. Meanwhile, RFI may also cause errors in spatial and radiometric measurements, and polarization and phase distortions may further affect the application effect of SAR images. From a signal bandwidth perspective, radio frequency interference can be divided into narrowband and wideband. With the development of modern communication technology, the bandwidth of ground unintentional interference sources is larger and larger, and the possibility of broadband radio frequency interference is greatly increased. And secondly, the broadband interference covers most of the frequency spectrum of the signal, the interference is difficult to distinguish from the signal, and the traditional suppression method is easy to damage useful signals, so that the suppression is more difficult. The SAR echo with broadband interference is processed, a useful target signal is protected, and the SAR echo is an important research direction facing the current complex electromagnetic environment. In order to solve the interference problem suffered by the SAR system, since 90 th of the last century, the scholars at home and abroad have conducted intensive research on SAR broadband interference suppression, and proposed various algorithms, which can be generally classified into parameterized, semi-parameterized and non-parameterized methods. The parameterization method mainly comprises the steps of establishing an accurate mathematical model of an interference signal, and further extracting the interference signal from echo data of SAR to achieve the inhibition effect. Such as maximum likelihood estimation, least squares estimation, parametric maximum likelihood and least mean square estimation algorithms, etc., are typical parameterized interference suppression methods. Liu Zhiling et al propose to estimate the model parameters of the narrowband interference by using an iterative adaptive method (IAA), obtain the interference suppression result considering both the speed and the efficiency, yang Zhiwei et al combine the time-frequency analysis technology, expand the IAA narrowband interference suppression method, and make it suitable for wideband interference suppression. The semi-parameterization method converts the complex signal separation problem into an optimization problem of super parameters for solving. From the angles of sparse reconstruction and low-rank decomposition theory, nguyen et al firstly put forward the concept of semi-parametric interference suppression, and the suppression of narrow-band interference is realized by solving the sparse reconstruction optimization problem, liu Hongqing et al expand the sparse reconstruction algorithm to wide-band interference, and provide a new thought for the expansion application of the method. Yellow rock et al propose a self-adaptive notch semi-parameter (ANSP) method, which combines the advantages of the semi-parameter method and a frequency domain notch filter (FNF) to obtain a better inhibition effect. The semi-parameterization method utilizes the optimization model to restrict interference, so that useful signals can be well protected, but the performance of the semi-parameterization method also depends on the value of the super-parameters and the selection of the optimization model, and the operation amount is much larger than that of other two methods. The non-parameterized method directly applies the energy characteristics of the interference signals to inhibi