MIMO-SAR echo separation method based on image post-processing in complex detection scene

Abstract

The invention discloses an MIMO-SAR echo separation method based on image post-processing in a complex detection scene, and belongs to the field of signal processing, and the method comprises the steps: designing an MIMO-SAR transmitted waveform, and collecting an echo signal through employing a pitching multi-channel system; carrying out distance direction and azimuth direction compression on pitching multichannel received data in sequence by adopting a classical imaging algorithm to complete imaging processing; constructing a pitching dimension airspace snapshot signal based on the multi-channel data after imaging processing, and selecting a distance direction sample to estimate a sample covariance matrix; eigenvalue decomposition is carried out on the covariance matrix to estimate the number of information sources in the airspace; under the condition that the number of the information sources is determined, the direction of arrival of the information sources is estimated through the MUSIC algorithm; according to the method, the optimal weight vector is calculated based on the least square criterion, the spatial domain snapshot is processed, interference signals in echoes are suppressed, and the problem of echo separation in MIMO-SAR is solved. According to the method, the influence of direction of arrival mismatch caused by topographic relief on the echo separation performance can be effectively relieved, and the imaging performance of the MIMO-SAR system in a complex detection scene is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of signal processing, and in particular to a MIMO-SAR echo separation method based on image post-processing in a complex detection scene. Background Art

[0002] Multiple-input multiple-output (MIMO) synthetic aperture radar (SAR) combines the MIMO principle with SAR technology, representing a major advancement in radar technology. In the past few decades, SAR has attracted widespread attention due to its ability to provide comprehensive global environmental monitoring. However, traditional SAR architectures that rely on single-transmitter-single-receiver or single-transmitter-multiple-receiver often encounter limitations in data acquisition speed, resolution, and adaptability to dynamic environments. To address the above issues, MIMO-SAR systems deploy multiple transmit and receive antennas, have virtual aperture expansion capabilities, and flexible imaging modes, significantly improving SAR capabilities in urban monitoring, environmental assessment, and disaster management.

[0003] Although the new MIMO-SAR has great potential, it also faces some challenges, such as the increased complexity of system design and data processing. In addition, due to the deployment characteristics of the transmitting and receiving antennas, the same receiver will receive the echo signals of all transmitted waveforms at the same time, while there are no transmitted waveforms with the same frequency and completely orthogonal at the same time. The resulting cross-correlation energy between waveforms will accumulate in the imaging scene, seriously deteriorating the imaging performance. How to effectively solve the echo separation problem in MIMO-SAR is crucial to the engineering application of MIMO-SAR systems.

[0004] In order to solve the echo separation problem in MIMO-SAR systems, a large number of theoretical studies and experimental verifications have been conducted in the existing literature. In order to provide more degrees of freedom in azimuth, the most popular MIMO-SAR echo separation scheme is the pitch-dimension beamforming echo separation scheme based on partially orthogonal waveforms. This scheme suppresses near-end interference through the partial orthogonal characteristics of the waveform, while far-end interference needs to be suppressed through the pitch-dimension DBF. However, the processing performance of this scheme depends on the DBF performance. In actual situations, there are often undulating terrain in MIMO-SAR detection scenes, which causes the actual DOA of the signal component to deviate from the theoretical value. The resulting DOA mismatch will seriously deteriorate the DBF processing performance. To deal with such problems, the traditional robust DBF technology uses various DOA estimation methods to estimate the incident angle of the signal component in the airspace. However, in the SAR detection process, the azimuth pulse expansion causes the DOA of the desired signal and the interference component in a certain unit to be processed to be not a uniquely determined value. At this time, the optimal weight vector calculated based on the DOA estimation result cannot apply a distortion-free response to the desired signal at all angles and fully suppress the interference component. Furthermore, existing literature has proposed to apply distortion-free response and nulling in all possible directions of the desired signal and interference components by widening the main lobe and nulling width of the beam pattern. This method can further improve the echo separation performance of MIMO-SAR under undulating terrain, but it consumes more system resources in the elevation dimension. When the MIMO-SAR system resources cannot meet the requirements, the processing performance of this method is also limited. Summary of the invention

[0005] The present invention provides a MIMO-SAR echo separation method based on image post-processing in complex detection scenes. The influence of azimuth pulse expansion on DOA mismatch is solved through image post-processing, and the incident angle of each signal component in the processing unit is estimated through the MUSIC algorithm, thereby alleviating the influence of terrain undulation in the detection scene on the MIMO-SAR echo separation performance. The method can avoid the influence of azimuth pulse expansion on DOA mismatch in complex detection scenes, simplify the signal model, and improve the DOA estimation accuracy and MIMO-SAR echo separation performance.

[0006] The embodiment of the present invention provides a MIMO-SAR echo separation method based on image post-processing in a complex detection scenario, comprising the following steps:

[0007] Step S1, designing the transmission waveform of the MIMO-SAR system and using a pitch multi-channel system to collect echo signals;

[0008] Step S2, using an imaging algorithm to sequentially compress the echo signals received by the elevation multi-channel system in the range and azimuth directions to complete the imaging process;

[0009] Step S3, constructing a pitch-dimensional spatial domain snapshot signal based on the multi-channel data after imaging processing, and selecting a sample covariance matrix for range-direction sample estimation;

[0010] Step S4, performing eigenvalue decomposition on the sample covariance matrix to estimate the number of information sources in the airspace;

[0011] Step S5, when the number of information sources is determined, the MUSIC algorithm is used to estimate the arrival angle of the information source;

[0012] Step S6, calculating the optimal weight vector based on the least squares criterion, and processing the spatial domain snapshot in step S3 to suppress interference signals in the echo.

[0013] Optionally, in one embodiment of the present invention, step S1 specifically includes:

[0014] A partially orthogonal waveform is used as the transmitting waveform of the MIMO-SAR system. m (t) is:

[0015]

[0016] Where t is the fast time, M is the number of transmitted waveforms, is the modulation vector, γ i is the modulation vector The element in the index is 1, () T is the transposition operation, m is the number of the transmitted waveform, i is the number of the sub-pulse, s0(t) is the signal basis representing a group of sub-pulses, T p is the pulse width;

[0017]

[0018]

[0019]

[0020] Where s(t) is a conventional linear frequency modulation signal, rect(t) is a rectangular window function, and the modulation vector is a unit vector, and the γth i elements are 1, to ensure that each transmitted waveform has the same bandwidth, the modulation vector satisfy:

[0021]

[0022] Where j is the sub-pulse number;

[0023] In the MIMO-SAR system, the same receiver will receive the echo signals of all transmitted waveforms. At this time, the echo signal r of the nth receiving channel isn (t,η) is expressed as:

[0024]

[0025] In the formula,

[0026] h n (t,η)=[h n,1 (t,η),h n,2 (t,η),…,h n,M (t,η)] T

[0027]

[0028] In the formula, h n,m (t,η) and r n,m (t,η) respectively represent the channel response and echo signal corresponding to the transmission of the mth transmitter and the reception of the nth receiver. and η are the fast time and slow time of convolution operation respectively.

[0029] Optionally, in one embodiment of the present invention, in step S2, imaging processing is performed using a range-Doppler algorithm, and range compression is performed using matched filtering.

[0030] Optionally, in one embodiment of the present invention, step S3 specifically includes:

[0031] When the MIMO-SAR system transmits M kinds of transmission waveforms simultaneously, there are at most 2M-1 signal components. At this time, the pitch-dimensional spatial domain snapshot is expressed as:

[0032]

[0033] In the formula, σ(θ q ) and v(θ q ) are the amplitude and steering vector of the signal component q, n is the noise steering vector, Represents echo data r n (t,η) imaging result, N is the number of receiving channels in the pitch dimension;

[0034] Sample covariance matrix It is expressed as:

[0035]

[0036] Where L is the number of distance samples selected, (·) H is the conjugate transpose operation, Take a quick shot of the airspace.

[0037] Optionally, in one embodiment of the present invention, step S4 specifically includes:

[0038] The sample covariance matrix after eigenvalue decomposition is:

[0039]

[0040] In the formula,

[0041] A=[v(θ0),v(θ1),…,v(θ K-1 )]

[0042] Λ S+I =diag[λ0,λ1,…,λ K-1 ]

[0043] Λ n =diag[λ K ,λ K+1 ,…,λ N-1 ]

[0044] In the formula, R S+I is the signal component covariance matrix, is the noise power, I is the unit matrix, K is the number of information sources, is the complex field, U c and Λ c is a matrix formed by eigenvectors and eigenvalues, U S+I and Λ S+I is the matrix formed by the signal eigenvectors and eigenvalues, U n and Λ n is the matrix formed by the noise eigenvector and eigenvalue, diag(·) is the operation to generate the diagonal matrix, λ k is the characteristic value;

[0045] The number of sources is calculated through the sample covariance matrix In the matrix U formed by the eigenvectors c The projection is estimated, that is:

[0046]

[0047] Define the detection factor:

[0048]

[0049] In the formula, u n is the projection vector, ||·||1 is the modulus operation, when β n When the maximum value appears, the corresponding subscript n is the number of information sources in the airspace.

[0050] Optionally, in one embodiment of the present invention, step S5 specifically includes:

[0051] The arrival angle estimation model based on the MUSIC algorithm is expressed as:

[0052]

[0053] In the formula, the set Represents all possible arrival angle values ​​of each signal component, for There are three signal components, and the corresponding arrival angle estimation value is expressed as:

[0054]

[0055] In the formula, is the angle of incidence.

[0056] Optionally, in one embodiment of the present invention, step S6 specifically includes:

[0057] When the incident angle of the signal component is θ q When , the signal steering vector is expressed as:

[0058]

[0059] μ p (θ)=2πpd sinθ / λ, p=0,1,…,N-1

[0060] In the formula, μ p (θ) is the phase part, p is the receiving array number, d and λ are the channel spacing and working wavelength respectively. The optimal weight vector is calculated using the least squares criterion. The specific calculation process is expressed as:

[0061]

[0062] Where W LS is the weighted vector matrix, using the weight vector Weighted processing is performed on the pitch-dimensional spatial domain snapshot signal to extract the incident angle The corresponding signal components are detected, while other signal components are suppressed, ultimately achieving MIMO-SAR echo separation.

[0063] The MIMO-SAR echo separation method based on image post-processing in complex detection scenarios of the embodiments of the present invention performs echo separation processing in the image domain, which can avoid the influence of azimuth pulse expansion on DOA mismatch, simplify the signal model, and improve the DOA estimation accuracy, thereby reducing the influence of terrain undulations in the detection scenario on the MIMO-SAR echo separation performance. It has good robustness and broad application prospects in actual processing.

[0064] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0066] Figure 1 This is a flow chart of a MIMO-SAR echo separation method based on image post-processing in a complex detection scenario according to an embodiment of the present invention;

[0067] FIG. 2( a ) is a diagram showing the relationship between the height and azimuth position of a point target at different slant ranges according to an embodiment of the present invention;

[0068] FIG2( b ) is a diagram showing the relationship between the angle of arrival error and the azimuth position of a point target at different slant ranges according to an embodiment of the present invention;

[0069] FIG. 3( a ) and FIG. 3( b ) respectively show the imaging performance in the range and azimuth directions of a point target 1 at a slant range of 10 kilometers according to an embodiment of the present invention;

[0070] FIG. 3( c ) and FIG. 3( d ) respectively show the imaging performance in the range and azimuth directions of a point target 2 at a slant range of 10 km according to an embodiment of the present invention;

[0071] FIG. 3( e ) and FIG. 3( f ) respectively show the imaging performance in the range and azimuth directions of a point target 3 at a slant range of 10 km according to an embodiment of the present invention;

[0072] FIG4( a ) is a terrain undulation model corresponding to a MIMO-SAR illumination scene according to an embodiment of the present invention;

[0073] FIG4( b ) is a diagram showing the angle of arrival errors corresponding to different positions under the terrain undulation model according to an embodiment of the present invention;

[0074] FIG5( a ) is a truth graph corresponding to a MIMO-SAR illumination scene according to an embodiment of the present invention;

[0075] FIG5( b ) is an imaging result of a mixed signal before echo separation according to an embodiment of the present invention;

[0076] FIG5( c ) is an echo separation result using LS beamforming technology under a traditional echo separation strategy according to an embodiment of the present invention;

[0077] FIG5( d ) is an echo separation result using the SMI beamforming technology under the traditional echo separation strategy of an embodiment of the present invention;

[0078] FIG5(e) is an echo separation result using the LCMV beamforming technology under the traditional echo separation strategy of an embodiment of the present invention;

[0079] FIG5( f ) is an echo separation result using the ESB beamforming technology under the traditional echo separation strategy of an embodiment of the present invention;

[0080] FIG5( g ) is an echo separation result using the Rec-Est beamforming technology under the traditional echo separation strategy of an embodiment of the present invention;

[0081] FIG5(h) is an echo separation result using LASSO beamforming technology under a traditional echo separation strategy according to an embodiment of the present invention;

[0082] FIG. 5( i ) is an imaging result after being processed by the proposed method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0083] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.

[0084] Figure 1 The present invention provides a flowchart of a MIMO-SAR echo separation method based on image post-processing in a complex detection scenario according to an embodiment of the present invention.

[0085] like Figure 1 As shown, the MIMO-SAR echo separation method based on image post-processing in the complex detection scenario includes the following steps:

[0086] Step S1, designing the transmission waveform of the MIMO-SAR system and using a pitch multi-channel system to collect echo signals.

[0087] The present invention realizes separation of MIMO-SAR echo signals based on partially orthogonal waveforms. The partially orthogonal waveforms are selected as the transmitting waveforms of the MIMO-SAR system, which can be regarded as consisting of linearly shifted or cyclically shifted sub-pulse signals. The transmitting waveform can be represented by a set of signal bases s0(t). The specific form of s0(t) can be expressed as:

[0088]

[0089]

[0090] The selected transmit waveform is a short-time shifted orthogonal (STSO) waveform. The general form of the MIMO-SAR partially orthogonal transmit waveform can be expressed as:

[0091]

[0092] Where, t, M, Tp , () T , s(t) and rect(t) represent fast time, number of transmitted waveforms, pulse width, transpose operation, conventional linear frequency modulation (LFM) signal and rectangular window function, respectively. s0(t) represents a signal basis composed of a set of sub-pulses, and the modulation vector is a unit vector, and the γth i elements are 1. In addition, to ensure that each transmitted waveform has the same bandwidth, Need to meet:

[0093]

[0094] For the transmission waveform m (t), the corresponding modulation matrix can be expressed as:

[0095]

[0096] In the formula, In order to ensure that the designed transmission waveform has partial orthogonal characteristics, the modulation matrix of different transmission waveforms must satisfy:

[0097] E i ⊙E j =0,1≤i,j≤M,i≠j

[0098] In the formula, 0 is a zero matrix. Considering that in the MIMO-SAR system, the same receiver receives the echo signals of all transmitters at the same time, the echo signal received by the receiver Rxn can be expressed as:

[0099]

[0100] In the formula,

[0101] h n (t,η)=[h n，1 (t,η),h n,2 (t,η),…,h n，M (t,η)] T

[0102]

[0103] In the formula, h n,m (t,η) and r n,m (t,η) respectively represent the channel response and echo signal corresponding to the transmission of the mth transmitter and the reception of the nth receiver. and η are the fast time convolution operation and the slow time respectively. n,m (t,η) is the channel response corresponding to transmitter m and receiver n, and its specific form is:

[0104] h n,m (t,η)=σ0δ(tt n,m )exp(-j2πf c t n,m )

[0105] Where, σ0, δ(t), t n,m and f c are the target amplitude, impulse function, target round-trip delay and carrier frequency respectively. By performing pulse compression processing on the echo signal, we can get:

[0106]

[0107] In the formula, () * is the conjugate operation, and are the desired signal component and the interference component respectively, and and β m,j The specific form is:

[0108]

[0109] Where κ{E(j,:)} will return the value of i that satisfies E(j,i)=1.

[0110] Step S2, using an imaging algorithm to sequentially compress the echo signals received by the elevation multi-channel system in the range and azimuth directions to complete the imaging process.

[0111] In this step, the range-Doppler algorithm (RDA) is selected for imaging processing, and range compression can be achieved through matched filtering. Before azimuth compression, due to azimuth pulse expansion, a unit to be processed will contain desired signals and interference components from different azimuths. Under the influence of terrain undulations, the desired signals and interference components at various azimuth positions will have different angles of arrival (DOA) values, namely:

[0112] Θ0θ0(α0), θ0(α1), θ0(α2),…

[0113] Θ q θ q (α0),θ q (α1),θ q (α2),…,-M+1≤q≤Ml,q≠0

[0114] In the formula, Θ0 and Θ q are the DOA sets of the desired signal and the interference component, θ0(α i ) and θ q (α i ) are the desired signal and interference component at azimuth angle α i DOA at.

[0115] After the echo data is compressed in azimuth, the DOA of the desired signal and the interference component in the unit to be processed can be approximately considered to have a unique definite value, which can effectively simplify the signal model and avoid the influence of azimuth pulse expansion on DOA mismatch.

[0116] This step uses the classic RDA algorithm to perform imaging processing on the echo data. In step S1, the echo data has been compressed in range. In this step, after completing the range migration correction, it will be further compressed in azimuth. The specific process will not be repeated here.

[0117] Step S3, constructing a pitch-dimensional spatial domain snapshot signal based on the multi-channel data after imaging processing, and selecting a sample covariance matrix for range-direction sample estimation.

[0118] Based on the transmit waveform designed in step S1, when the MIMO-SAR system transmits M transmit waveforms simultaneously, there can be at most 2M-1 signal components in the unit to be processed. At this time, the pitch-dimensional spatial snapshot can be expressed as:

[0119]

[0120] In the formula, σ(θ q ) and v(θ q ) are the amplitude of the signal component q, the steering vector, n is the noise steering vector, Represents echo data r n The imaging result of (t,η), N is the number of receiving channels in the pitch dimension. In general, the sample covariance matrix can be expressed as:

[0121]

[0122] In the formula, L and (·) H are the selected distance direction sample number and conjugate transpose operation respectively.

[0123] Step S4, performing eigenvalue decomposition on the sample covariance matrix to estimate the number of information sources in the spatial domain.

[0124] After eigenvalue decomposition, the covariance matrix can be further expressed as:

[0125]

[0126] In the formula,

[0127] A=[v(θ0),v(θ1),…,v(θ K-1 )]

[0128] Λ S+I =diag[λ0,λ1,…,λ K-1 ]

[0129] Λ n =diag[λ K ,λ K+1 ,…,λ N-1 ]

[0130] In the formula, R S+I , I, K and are the signal component covariance matrix, noise power, identity matrix, number of sources and complex domain, U c and Λ c is a matrix formed by eigenvectors and eigenvalues, U S+I and Λ S+I is the matrix formed by the signal eigenvectors and eigenvalues, U n and Λ n is the matrix formed by the noise eigenvectors and eigenvalues, and diag(·) is the operation to generate a diagonal matrix.

[0131] In order to reduce the impact of noise on the estimation of the number of information sources, the sample covariance can be subjected to noise reduction processing, namely:

[0132]

[0133] In the formula, γ 2 The noise estimation power can be obtained by estimating the small eigenvalue generated by eigenvalue decomposition. Furthermore, the above formula can be further expressed as:

[0134]

[0135] In the formula,

[0136] Λ′ S+I =diag[λ0-γ 2 ,λ1-γ 2 ,…,λ K-1 -γ 2 ]

[0137] Λ′ n =diag[λ K -γ 2 ,λ K+1 -γ 2 ,…,λ N-1 -γ 2 ]

[0138] In this step, the number of sources is based on R′ in U c The projection is estimated by , and the projection can be calculated as:

[0139]

[0140] Define the detection factor:

[0141]

[0142] In the formula, ||·||1 is the modulo operation. n When the maximum value appears, the corresponding subscript n is the number of information sources.

[0143] Furthermore, the number of information sources in the airspace can be estimated as:

[0144]

[0145] In the formula,

[0146] =[1,β2,…,β N-1 ] T .

[0147] Step S5: when the number of information sources is determined, the MUSIC algorithm is used to estimate the arrival angle of the information source.

[0148] When there is terrain undulation, the incident angle of each signal component cannot be determined based on prior information, so DOA estimation is required through the sample covariance matrix. The DOA estimation model based on the MUSIC algorithm can be expressed as:

[0149]

[0150] In the formula, the set Represents all possible DOA values ​​of each signal component. There are three signal components, and the corresponding DOA estimation value can be expressed as:

[0151]

[0152] Step S6, calculate the optimal weight vector based on the least squares criterion, and process the spatial domain snapshot in step S3 to suppress the interference signal in the echo and solve the echo separation problem in MIMO-SAR.

[0153] When the incident angle of the signal component is θ q ′, the signal steering vector can be expressed as:

[0154]

[0155] μ p (θ)=2πpd sinθ / λ,p=0,1,…,N-1

[0156] In the formula, d and λ are the channel spacing and the working wavelength respectively. This step uses the least squares criterion to calculate the optimal weight vector. The specific calculation process can be expressed as:

[0157]

[0158] Using weight vector By weighting the pitch-dimensional spatial snapshots, the incident angle can be extracted The corresponding signal components are detected, while other signal components are suppressed, ultimately achieving MIMO-SAR echo separation.

[0159] Furthermore, simulation data is used to verify the processing performance of the MIMO-SAR echo separation method based on image post-processing in a complex detection scenario proposed by the present invention. The simulation experiment parameters involved in the experiment are shown in Table 1.

[0160] Table 1 Main parameters involved in simulation data

[0161]

[0162] The MIMO-SAR system will transmit two transmission waveforms simultaneously, and the four channels in the elevation dimension will receive echo signals simultaneously. The specific forms of the two transmission waveforms are:

[0163]

[0164] In the formula, k r Represents the frequency modulation slope. First, the echo separation performance of the proposed method is verified through point target simulation experiments. Figure 2(a) shows the relationship between the point target height at 9.5km, 10km and 10.5km and the azimuth position of the point target, and Figure 2(b) shows the relationship between the point target arrival angle error and the azimuth position at different positions. In actual processing, the true height information of each point target cannot be known a priori, so there will be DOA mismatch in the multi-channel echo signal. After using the proposed method for echo separation processing, three point targets at 10 kilometers are selected for imaging performance analysis. The imaging results of the three point targets are shown in Figure 2(a). Figure 3(a)-Figure 3(f) As shown in the figure, it can be seen that the range imaging performance of each point target does not change before and after echo separation, so the waveform cross-correlation energy does not affect the range imaging performance of the target. In addition, after echo separation, the azimuth imaging performance of the point target is significantly improved, which also confirms the effectiveness of the proposed method.

[0165] Furthermore, a surface target simulation experiment is used to verify the performance of the proposed method. Figure 4(a) shows the terrain undulation model of the MIMO-SAR illumination scene, and its height variation range is -150m to 150m. The range dimension of the imaging scene is 1.5 kilometers, and the azimuth dimension is 1.2 kilometers. Figure 4(b) shows the angle of arrival mismatch model corresponding to the terrain undulation model. In order to highlight the superiority of the proposed method, the traditional least squares (LS) beamforming technology, sample matrix inversion (SMI) beamforming technology, linear constrained minimum variance (LCMV) beamforming technology, characteristic subspace (ESB) beamforming technology, covariance matrix reconstruction (Rec-Est) beamforming technology, and least absolute shrinkage and selection operator (LASSO) beamforming technology are used to separate the echo data. Figure 5(a)-Figure 5(i) Indicates the processing results of different methods. Figure 5(a) is the true value image of the illuminated scene, and Figure 5(b) is the imaging result of the mixed signal before echo separation. It can be seen that the cross-correlation energy between waveforms will seriously deteriorate the imaging quality of MIMO-SAR. Figure 5(c)-Figure 5(i) These are the echo separation results of LS, SMI, LCMV, ESB, Rec-Est, LASSO and the proposed technology. It should be noted that the LS, SMI, LCMV, ESB, Rec-Est and LASSO beamforming technologies perform echo separation processing before azimuth compression. From the processing results, it can be seen that LASSO and the proposed method can obtain better processing performance. In order to quantitatively analyze the echo separation performance of different methods, the structural similarity (SSIM) index is used to quantitatively evaluate the four regions in the imaging results. Generally speaking, the higher the calculated SSIM value, the better the processing performance of the method. By comparison, the four regions of the proposed method all have the highest SSIM values, which further proves the superiority of the proposed method and has broad application prospects in MIMO-SAR data processing.

[0166] The MIMO-SAR echo separation method based on image post-processing in complex detection scenarios proposed in an embodiment of the present invention can effectively alleviate the impact of the angle of arrival mismatch caused by terrain undulation on the echo separation performance, and improve the imaging performance of the MIMO-SAR system in complex detection scenarios (with terrain undulations).

[0167] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or N embodiments or examples in a suitable manner. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0168] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "N" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0169] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or N executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present invention includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present invention belong.

Claims

1. A MIMO-SAR echo separation method based on image post-processing in complex detection scenarios, characterized in that: The following steps are involved: Step S1, designing the transmission waveform of the MIMO-SAR system and using a pitch multi-channel system to collect echo signals; Step S2, using an imaging algorithm to sequentially compress the echo signals received by the elevation multi-channel system in the range and azimuth directions to complete the imaging process; Step S3, constructing a pitch-dimensional spatial domain snapshot signal based on the multi-channel data after imaging processing, and selecting a sample covariance matrix for range-direction sample estimation; Step S4, performing eigenvalue decomposition on the sample covariance matrix to estimate the number of information sources in the airspace; Step S5, when the number of information sources is determined, the MUSIC algorithm is used to estimate the arrival angle of the information source; Step S6, calculating the optimal weight vector based on the least squares criterion, and processing the spatial domain snapshot in step S3 to suppress interference signals in the echo.

2. The method according to claim 1, characterized in that Step S1 specifically includes: A partially orthogonal waveform is used as the transmitting waveform of the MIMO-SAR system. m (t) is: Where t is the fast time, M is the number of transmitted waveforms, is the modulation vector, γ i is the modulation vector The element in the index is 1, () T is the transposition operation, m is the number of the transmitted waveform, i is the number of the sub-pulse, s0(t) is the signal basis representing a group of sub-pulses, T p is the pulse width; s0(t)=[w1(t)s(t),w2(t)s(t),…,w M (t)s(t)] T Where s(t) is a conventional linear frequency modulation signal, rect(t) is a rectangular window function, and the modulation vector is a unit vector, and the γth i elements are 1, to ensure that each transmitted waveform has the same bandwidth, the modulation vector satisfy: Where j is the sub-pulse number; In the MIMO-SAR system, the same receiver will receive the echo signals of all transmitted waveforms. At this time, the echo signal r of the nth receiving channel is n (t,η) is expressed as: In the formula, h n (t,η)=[h n,1 (t,h),h n,2 (t,h),…,h n,M (t,n)] T In the formula, h n,m (t,η) and r n,m (t,η) respectively represent the channel response and echo signal corresponding to the transmission of the mth transmitter and the reception of the nth receiver. and η are the fast time and slow time of convolution operation respectively.

3. The method according to claim 1, characterized in that In step S2, imaging processing is performed using a range-Doppler algorithm, and range compression is performed using matched filtering.

4. The method according to claim 3, characterized in that Step S3 specifically includes: When the MIMO-SAR system transmits M kinds of transmission waveforms simultaneously, there are at most 2M-1 signal components. At this time, the pitch-dimensional spatial domain snapshot is expressed as: In the formula, σ(θ q ) and v(θ q ) are the amplitude and steering vector of the signal component q, n is the noise steering vector, Represents echo data r n (t,η) imaging result, N is the number of receiving channels in the pitch dimension; Sample covariance matrix It is expressed as: Where L is the number of distance samples selected, (·) H is the conjugate transpose operation, χ l Take a quick shot of the airspace.

5. The method according to claim 4, characterized in that Step S4 specifically includes: The sample covariance matrix after eigenvalue decomposition is: In the formula, A=[v(θ0),v(θ1),…,v(θ K-1 )] L S+I =diag[λ0,λ1,…,λ K-1 ] L n =diag[λ K ,l K+1 ,…,l N-1 ] In the formula, R S+I is the signal component covariance matrix, is the noise power, I is the unit matrix, K is the number of information sources, is the complex field, U c and Λ c is a matrix formed by eigenvectors and eigenvalues, U S+I and Λ S+I is the matrix formed by the signal eigenvectors and eigenvalues, U n and Λ n is the matrix formed by the noise eigenvector and eigenvalue, diag(·) is the operation to generate the diagonal matrix, λ k is the characteristic value; The number of sources is obtained through the sample covariance matrix In the matrix U formed by the eigenvectors c The projection is estimated, that is: Define the detection factor: In the formula, u n is the projection vector, ||·||1 is the modulus operation, when β n When the maximum value appears, the corresponding subscript n is the number of information sources in the airspace.

6. The method according to claim 5, characterized in that Step S5 specifically includes: The arrival angle estimation model based on the MUSIC algorithm is expressed as: In the formula, the set Represents all possible angle of arrival values ​​of each signal component, for There are three signal components, and the corresponding arrival angle estimation value is expressed as: In the formula, is the angle of incidence.

7. The method according to claim 1, characterized in that Step S6 specifically includes: When the incident angle of the signal component is θ q When , the signal steering vector is expressed as: m p (θ)=2πpdsinθ / λ,p=0,1,…,N-1 In the formula, μ p (θ) is the phase part, p is the receiving array number, d and λ are the channel spacing and working wavelength respectively. The optimal weight vector is calculated using the least squares criterion. The specific calculation process is expressed as: Where W LS is the weighted vector matrix, using the weight vector Weighted processing is performed on the pitch-dimensional spatial domain snapshot signal to extract the incident angle The corresponding signal components are detected, while other signal components are suppressed, ultimately achieving MIMO-SAR echo separation.

Images