Complex target radar forward-looking imaging method based on sparse representation and GGAMP-VSBL algorithm

Abstract

The application provides a complex target radar forward-looking imaging method based on sparse representation and a GGAMP-VSBL algorithm, a radiation field reference signal and a model architecture of a forward-looking imaging model are determined according to imaging requirements; a random frequency modulation signal is transmitted according to the model architecture, and a target echo signal is received by a receiving array element, and the target echo signal can be expressed as superposition of the radiation field reference signal in an imaging plane; then an expression form of the forward-looking imaging model is obtained; the target echo signal and the radiation field reference signal are taken as known conditions, and are substituted into the forward-looking imaging model to obtain a target scattering coefficient, and finally a target image is reconstructed. The application firstly performs sparse transformation on the target, and changes solving a complex target into solving a sparse target, so that a better result is obtained. Since energy is basically concentrated on an imaging target point and there is no false scattering point, the application can realize recovery of the target, and compared with other methods, the performance is superior, and the application has a strong ability of recovering the target.

Description

Technical Field

[0001] This invention belongs to the field of radar imaging technology, specifically relating to a forward-looking radar imaging method for complex targets based on sparse representation and the GGAMP-VSBL algorithm. Background Technology

[0002] Forward-looking radar imaging, as a crucial component of radar imaging technology, has been a cutting-edge and hotly debated area of ​​radar research in recent years. As an important method for acquiring target information, forward-looking radar imaging plays a vital role in target estimation, forward-looking detection, and precision guidance. With increasing demands for radar imaging accuracy and the complexity of imaging targets, achieving high-resolution imaging of complex targets has become a key technical challenge that needs to be addressed in the imaging process.

[0003] Traditional inverse synthetic aperture radar (ISAR) and synthetic aperture radar (SAR) technologies utilize the relative motion between the radar and the target to achieve high-resolution imaging by differentiating the Doppler frequency shift of the scattering points. SAR is typically mounted on moving platforms such as aircraft and satellites, acquiring target images through radar motion. In contrast, in ISAR imaging, the radar is generally stationary, and imaging is achieved by utilizing the target's motion relative to the radar. Real aperture imaging and phased array radar imaging technologies transmit periodic, fixed signal waveforms. Within the beam coverage area, the radiated wavefront is essentially the same at different times. The receiving array elements receive the scattered echoes from the target and process them to obtain the target's image.

[0004] Traditional inverse synthetic aperture radar (ISAR) and synthetic aperture radar (SAR) technologies suffer from Doppler ambiguity and forward-looking imaging blind zones. Furthermore, these methods require processing large amounts of data, resulting in high computational complexity. They also rely on the relative motion between the target and the radar for imaging. Real aperture imaging and phased array radar imaging technologies are limited by antenna size, leading to weaker imaging capabilities when the antenna size is small. In addition, as the number of radiations increases during imaging, periodic scattering echoes cannot provide additional information on the resolution of targets within the beam; they can only improve the signal-to-noise ratio. Therefore, it is extremely difficult to distinguish targets within the antenna beam solely by relying on the beam, resulting in unsatisfactory imaging effects. Summary of the Invention

[0005] To address the aforementioned problems in the existing technology, this invention provides a forward-looking imaging method for complex targets based on sparse representation and the GGAMP-VSBL algorithm. The technical problem to be solved by this invention is achieved through the following technical solution:

[0006] This invention provides a forward-looking radar imaging method for complex targets based on sparse representation and the GGAMP-VSBL algorithm, comprising:

[0007] S100, acquire imaging requirements, which include radar imaging requirements and target imaging requirements;

[0008] S200, determine the radiation field reference signal and the model architecture of the forward-looking imaging model according to the imaging requirements;

[0009] S300, a random frequency modulated signal is transmitted according to the model architecture, and the target echo signal is obtained based on the superposition relationship between the echo signal and the radiation field reference signal in the imaging plane;

[0010] S400, the expression form of the forward-looking imaging model is obtained according to the relationship between the target echo signal, the target scattering coefficient, the radiation reference signal, and the model architecture;

[0011] S500, the target echo signal and the radiation field reference signal are taken as known conditions and substituted into the forward-looking imaging model to obtain the target scattering coefficient; wherein, the target scattering coefficient is obtained by sparsification solution;

[0012] S600, the target image is reconstructed based on the target scattering coefficient.

[0013] Beneficial effects:

[0014] This invention provides a forward-looking radar imaging method for complex targets based on sparse representation and the GGAMP-VSBL algorithm. The method determines the radiation field reference signal and the model architecture of the forward-looking imaging model based on imaging requirements. A random frequency-modulated signal is transmitted according to the model architecture, and the echo signal is received using receiving array elements. The target echo can be represented as the superposition of the radiation field reference signal in the imaging plane. Then, the expression form of the forward-looking imaging model is obtained. The target echo signal and the radiation field reference signal are used as known conditions and substituted into the forward-looking imaging model to obtain the target scattering coefficients. Finally, the target image is reconstructed. This invention first performs a sparse transformation on the target, changing the solution from a complex target to a sparse target, thus obtaining better results. Since the energy is basically concentrated on the imaging target point, there are no false scattering points. This invention can achieve target recovery, exhibiting superior performance compared to other methods and possessing a strong ability to recover targets.

[0015] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the radar forward-looking imaging model of the present invention;

[0017] Figure 2 This is a flowchart of the complex target radar forward-looking imaging method based on sparse representation and GGAMP-VSBL algorithm of the present invention;

[0018] Figure 3 Comparison of digital target simulation imaging results;

[0019] Figure 4 Comparison of simulation imaging results for aircraft surface targets. Detailed Implementation

[0020] The present invention will be further described in detail below with reference to specific embodiments, but the implementation of the present invention is not limited thereto.

[0021] Combination Figures 1 to 4 As shown, this invention provides a forward-looking radar imaging method for complex targets based on sparse representation and the GGAMP-VSBL algorithm, comprising:

[0022] S100, acquire imaging requirements, which include radar imaging requirements and target imaging requirements;

[0023] This invention uses target detection and target tracking technology to estimate the center position of the target before imaging, and uses this as the center to determine the imaging plane. It only needs to focus on the radar signal within the imaging plane and divides the imaging plane into several grid cells of the same size.

[0024] refer to Figure 1 As shown, assuming a radar array contains N transmitting antennas and 1 receiving antenna, with each transmitting element emitting an independent random frequency modulated signal, a model architecture for a forward-looking imaging model of a complex target on the imaging plane is established. Without loss of generality, if a two-dimensional imaging problem is considered, the radar imaging requirements limit the number of transmitting and receiving elements, as well as the form of the random frequency modulated signal emitted by the transmitting elements; the target imaging requirements limit the grid division method and the number of grids on the target imaging plane.

[0025] S200, determine the radiation field reference signal and the model architecture of the forward-looking imaging model according to the imaging requirements;

[0026] Assume there are N transmitting elements and 1 receiving element. Each transmitting element transmits an independent random frequency modulated signal. The random frequency modulated signal transmitted by the nth transmitting element is St. n (t), the position vector of the transmitting array element is R n The target echo signal received by the receiving array element is S R (t), the position vector of the receiving array element is R0, c is the propagation speed of electromagnetic waves in the atmosphere, and the random frequency-modulated signal of N transmitting array elements on the l-th imaging plane grid is represented as:

[0027]

[0028] Where t represents time.

[0029] The signal at each imaging unit grid is received by the receiving antenna unit after passing through the scattering coefficient, forming a radiation field reference signal. The radiation field reference signal formed after scattering from the l-th grid is expressed as:

[0030]

[0031]

[0032] Where, τ n,0 This represents the time delay of the random frequency modulated signal transmitted by the nth transmitting element after being scattered by the scattering point and reaching the receiving element.

[0033] S300, a random frequency modulated signal is transmitted according to the model architecture, and the target echo signal is obtained based on the superposition relationship between the echo signal and the radiation field reference signal in the imaging plane;

[0034] The received signal can be represented as a superposition of reference signals of the radiation field within the imaging plane. Compared to the detection signal, the radiation field reference signal only adds the time delay from the scattering point to the receiving array element. Therefore, the target echo signal can be obtained using the radiation field reference signal.

[0035]

[0036] Where, β l Let be the scattering coefficient of the target scattering point located at the l-th imaging plane grid. The target scattering coefficient vector can be expressed as β=[β1,β2,…,β L ] T .

[0037] If the target echo signal is sampled M times, the sampled target echo signal is represented as follows:

[0038] S R (t)=[S R (t0),S R (t1),…,S R (t M-1 )] T (5);

[0039] S400, the expression form of the forward-looking imaging model is obtained according to the relationship between the target echo signal, the target scattering coefficient, the radiation reference signal, and the model architecture;

[0040] The forward-looking imaging model can be represented in the following matrix form:

[0041]

[0042] The vector form of the forward-looking imaging model is as follows:

[0043] SR =S·β+n (7);

[0044] Where S R Let represent the target echo signal vector, S represent the radiation field reference matrix, β represent the target scattering coefficient vector, and n represent the noise vector.

[0045] S500, the target echo signal and the radiation field reference signal are taken as known conditions and substituted into the forward-looking imaging model to obtain the target scattering coefficient; wherein, the target scattering coefficient is obtained by sparsification solution;

[0046] Specifically, S500 of the present invention includes:

[0047] S510, the target scattering coefficient is expressed as an expression of the sparse representation coefficients multiplied by the transformation matrix;

[0048] For imaging complex targets, sparsity can be obtained through a certain transformation. Assume there exists a transformation matrix D such that the complex target β has a sparse representation σ, i.e.

[0049] β=Dσ (8);

[0050] Where σ is the sparse representation coefficient and D is the transformation matrix;

[0051] S520 uses the expression in S510 to convert the vector representation of the forward-looking imaging model into a sparse representation. After applying the sparse representation method, the vector representation of the forward-looking imaging model is expressed as the following sparse expression:

[0052] S R =SDσ+n (9);

[0053] S530, the target echo signal and the radiation field reference signal are used as known conditions, and the sparse representation coefficients are estimated using the GGAMP-VSBL algorithm;

[0054] Let S′ = SD, then the predicted sparse representation coefficients are:

[0055]

[0056] Where η represents the regularization parameter related to the noise level, and S′=SD;

[0057] S540, substitute the estimated sparse representation coefficients into the expression in S510 to estimate the target scattering coefficients. The estimated target scattering coefficients are then expressed as:

[0058]

[0059] S600, the target image is reconstructed based on the target scattering coefficient.

[0060] The target scattering coefficient vector is obtained using the GGAMP-VSBL algorithm. The image of the target can then be reconstructed. As can be seen from the above steps, σ, as the sparse coefficient vector of the sparse representation, has an important influence on the imaging result. The key is to find the transformation matrix D so that the complex target has a sparse representation. This invention uses a regularized approximate K-Singular Value Decomposition (AK-SVDr) dictionary learning algorithm to perform sparse representation of complex targets.

[0061] The radar forward-looking imaging process can be derived from the above derivation as follows: Figure 2 As shown, before imaging, the imaging plane needs to be gridded. A radiation field reference signal is obtained by transmitting a signal and using the imaging plane. The target echo is received by the receiving array element, the two are correlated, and the target image is obtained by the target reconstruction algorithm.

[0062] Based on the above description of the forward-looking imaging steps and methods, in order to verify the superior performance of the proposed method for imaging complex targets, simulation experiments were conducted using handwritten digits and aircraft surface targets as examples. The handwritten digit targets were derived from a handwritten character dataset. The typical characteristic of complex targets compared to sparse targets is that the scattering points are clustered together in a relatively concentrated position. This weakens the algorithm's ability to distinguish scattering points and affects the final imaging effect. However, if the complex targets are first sparsified and the algorithm is transformed into solving sparse targets, the final imaging quality will be greatly improved.

[0063] To demonstrate the effectiveness of the proposed algorithm in imaging complex targets and to verify its superior performance, the main simulation parameter settings are first presented in Table 1.

[0064] Table 1 Simulation parameter settings

[0065] Simulation parameters numerical values unit signal carrier frequency 10 GHz signal bandwidth 1 GHz Random radiation number 2000 / Number of antenna array elements 17 / Array element spacing 1 m Grid cell size 1×1 <![CDATA[m 2 ]]> Signal-to-noise ratio 15 dB Imaging plane distance 100 m

[0066] In this invention, we conducted simulation experiments to verify the effectiveness of the sparse representation method for imaging complex targets. The imaging plane was 32m × 32m in size, and was uniformly divided into 32 × 32 grids, with each grid measuring 1m × 1m. Specific experimental results are as follows: Figure 3 and Figure 4 As shown, Figure 3 Figure a in the middle and Figure 4 Figure a in the image shows the original image target scene. Figure 3 Figure b in the middle and Figure 4 Image b in the image shows the imaging results using the SBL algorithm. Figure 3 Figure c in the middle and Figure 4 In this context, 'c' represents the Tikhonov regularization method. Figure 3 d-graph and Figure 4 Figure d in the diagram illustrates the method of this invention. From... Figure 3 It can be seen that the Tikhonov regularization method is greatly affected by noise, and can only distinguish the general outline of the target. The imaging results contain many false scattering points and some target scattering points are missing. The SBL algorithm improves the imaging effect compared to the Tikhonov regularization method, and the target amplitude is enhanced to a certain extent. The method of this invention first performs a sparse transformation on the target, changing from solving a complex target to solving a sparse target, thus obtaining better results. The energy is basically concentrated on the imaging target points, with no false scattering points, and the target is basically restored. Compared with other methods, it has superior performance and a strong ability to restore the target. Figure 4 For aircraft surface targets, since the sparse prior is no longer satisfied, SBL cannot reconstruct the target, while Tikhonov regularization can only reconstruct the target outline. However, the method of this invention can not only see the target outline, but also the scattering points are all concentrated at the target position. As can be seen from the imaging results, this invention can better recover the target scattering points and thus obtain the best imaging results.

[0067] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0068] Although this application has been described herein in conjunction with various embodiments, those skilled in the art will understand and implement other variations of the disclosed embodiments by reviewing the accompanying drawings, the disclosure, and the appended claims in carrying out the claimed application. In the claims, the word "comprising" does not exclude other components or steps, and "a" or "an" does not exclude a plurality.

[0069] The above description, in conjunction with specific preferred embodiments, provides a further detailed explanation of the present invention. It should not be construed that the specific implementation of the present invention is limited to these descriptions. For those skilled in the art, various simple deductions or substitutions can be made without departing from the concept of the present invention, and all such modifications and substitutions should be considered within the scope of protection of the present invention.

Claims

1. A forward-looking radar imaging method for complex targets based on sparse representation and the GGAMP-VSBL algorithm, characterized in that, include: S100, acquire imaging requirements, which include radar imaging requirements and target imaging requirements; S200, based on the imaging requirements, determine the radiation field reference signal and the model architecture of the forward-looking imaging model, including N transmitting elements and 1 receiving element, with each transmitting element transmitting an independent random frequency-modulated signal. The random frequency modulated signal transmitted by each transmitting element is The position vector of the transmitting array element is The target echo signal received by the receiving array element is The position vector of the receiving array element is , Let be the speed of electromagnetic wave propagation in the atmosphere, and let the position vector at the grid be . r l ; in the The random frequency-modulated signal of N transmission elements on an imaging plane grid is represented as: （1）； in, t Indicates time; No. The reference signal of the radiation field formed after scattering by each grid is represented as: （2）； （3）； in, Represented as the first The time delay between the random frequency modulated signal emitted by each transmitting element and the scattering point before it reaches the receiving element; S300, a random frequency modulated signal is transmitted according to the model architecture, and the target echo signal is obtained based on the superposition relationship between the echo signal and the radiation field reference signal in the imaging plane; S400, the forward-looking imaging model is expressed according to the relationship between the target echo signal, the target scattering coefficient, the radiation field reference signal, and the model architecture; S500, the target echo signal and the radiation field reference signal are taken as known conditions and substituted into the forward-looking imaging model to obtain the target scattering coefficient; wherein, the target scattering coefficient is obtained by sparsification solution; S600, the target image is reconstructed based on the target scattering coefficient; Prior to S500, the complex target radar forward-looking imaging method also included sampling the target echo signal M times to obtain the sampled target echo signal, which is represented as: （5）； The forward-looking imaging model in S500 is represented as follows: （6） in, It is located at the th The scattering coefficients of the target scattering points at each imaging plane grid are expressed as: The target scattering coefficient vector is... ; The vector form of the forward-looking imaging model is as follows: （7）； in Represents the target echo signal vector. Represents the radiation field reference matrix. Represents the target scattering coefficient vector. Represents the noise vector; The S500 includes: S510, the target scattering coefficient is expressed as an expression of the sparse representation coefficients multiplied by the transformation matrix; S520 uses the expression in S510 to convert the vector of the forward-looking imaging model into a sparse representation. S530, the target echo signal and the radiation field reference signal are used as known conditions, and the sparse representation coefficients are estimated using the GGAMP-VSBL algorithm; S540, substitute the estimated sparse representation coefficients into the expression in S510 to estimate the target scattering coefficients.

2. The method for forward-looking radar imaging of complex targets based on sparse representation and GGAMP-VSBL algorithm according to claim 1, characterized in that, The radar imaging requirements define the number of transmitting and receiving array elements, as well as the form in which the transmitting array elements transmit random frequency modulated signals; the target imaging requirements define the grid division method of the target imaging plane and the number of grids in the imaging plane.

3. The method for forward-looking radar imaging of complex targets based on sparse representation and GGAMP-VSBL algorithm according to claim 1, characterized in that, The target echo signal in S300 is represented as follows: （4）。 4. The complex target radar forward-looking imaging method based on sparse representation and GGAMP-VSBL algorithm according to claim 3, characterized in that, The expression in S510 is represented as follows: （8）； in, For sparse representation coefficients, Represents the transformation matrix; The vector representation of the forward-looking imaging model in S520 is expressed as a sparse expression as follows: （9）； The estimated sparse representation coefficients are: （10）； in, This represents a regularization parameter related to the noise level. ; The target scattering coefficient predicted in S540 is expressed as follows: （11）。

Images