Fast range super-resolution imaging method based on GNSS-R SAR

Abstract

The present invention belongs to the technical field of GNSS-R SAR super-resolution imaging, and discloses a fast range super-resolution imaging method based on GNSS-R SAR. The present invention innovatively combines matrix compression, regularization modeling and fast optimization algorithm to form a set of efficient and stable range super-resolution processing procedures. Specifically, after obtaining the preliminary imaging results of the GNSS echo signal and giving its signal convolution model form, the observation matrix in the signal convolution model is first reduced in dimension by singular value decomposition, and on the basis of the processing results of the weighted singular value preservation strategy, the signal convolution model is reconstructed using the inverse matrix of the truncated measurement matrix; then, starting from the regularization strategy, the L1 norm constraint is introduced to construct the objective function using the sparse characteristics of the target; finally, a fast iterative optimization algorithm is used to solve the objective function. The method of the present invention significantly improves the range resolution of GNSS echo data.

Description

Technical Field

[0001] The present invention belongs to the technical field of GNSS-R SAR super-resolution imaging, and in particular relates to a fast range super-resolution imaging method based on GNSS-R SAR. Background Art

[0002] The Global Navigation Satellite System Reflectometry Synthetic Aperture Radar (GNSS-R SAR) system, based on navigation satellite backscatter signals, is an integrated space-ground, dual- or multi-static SAR system composed of navigation satellites, low-orbit satellites, and airborne or ground-based receivers. This system fully utilizes existing navigation satellite resources and offers advantages such as a large number of satellites, flexible and diverse geometric configurations, high concealment, short revisit periods, and long observation times. It is a key development direction for future space-ground, space-ground SAR radar networks.

[0003] Dual- and multistatic SAR systems, based on navigation satellites, utilize existing navigation satellite signals as external radiation sources. Receivers are placed on low-orbit satellites, airborne platforms, and fixed stations on the ground to receive reflected and scattered signals from the observation area and perform SAR imaging signal processing. Because GNSS signals are designed for positioning rather than imaging, they have low signal power and narrow bandwidth, resulting in low resolution and significant noise in radar imaging.

[0004] Super-resolution imaging technology primarily addresses the physical hardware limitations of traditional imaging systems (such as sensor bandwidth, antenna size, and the optical diffraction limit). Through advanced signal processing and reconstruction algorithms, it extracts or restores image details higher than the sampled resolution. Super-resolution imaging typically treats a low-resolution image as a high-resolution image that has been blurred (e.g., by the point spread function), downsampled, and subjected to noise. By accurately modeling this degradation, deconvolution or reconstruction algorithms can be designed to recover greater detail.

[0005] At present, since the bandwidth of the GPS ranging code is 1.023MHZ, its range resolution is about 150m. When the distance between two objects is less than 150m, traditional imaging methods cannot distinguish the two targets, so super-resolution imaging is required. Summary of the Invention

[0006] The purpose of the present invention is to propose a fast range super-resolution imaging method based on GNSS-R SAR for GNSS echo signals. The method combines matrix compression, regularized modeling and fast optimization algorithm to form an efficient and stable range super-resolution processing flow, so as to significantly improve the range resolution of GNSS-R imaging.

[0007] In order to achieve the above-mentioned purpose, the present invention adopts the following technical solutions:

[0008] A fast range super-resolution imaging method based on GNSS-R SAR includes the following steps:

[0009] Step 1. Obtain GNSS echo signals and perform preliminary imaging of the GNSS echo signals using the range-Doppler algorithm. The preliminary imaging results are expressed as a signal convolution model.

[0010] The observation matrix in the convolution model is the circulant matrix of the normalized autocorrelation function of the GNSS ranging code;

[0011] Step 2. Use singular value decomposition to reduce the dimension of the measurement matrix. At the same time, construct a weight function and propose a weighted singular value preservation strategy to reconstruct the measurement matrix in a weighted manner, thereby constructing a new signal convolution model.

[0012] Step 3. Based on the new signal convolution model, starting from regularization, taking advantage of the sparse nature of the target, and using the L1 norm as a constraint term to construct the objective function, the imaging problem is transformed into an inverse problem of sparse reconstruction.

[0013] Step 4. Design an accelerated iterative optimization algorithm to solve the objective function. The iterative optimization algorithm performs gradient descent combined with soft threshold shrinkage in each iteration, and introduces a momentum term to accelerate convergence.

[0014] Step 5. After the objective function is solved, the fast range super-resolution imaging results based on GNSS-R SAR are output.

[0015] In addition, based on the fast range super-resolution imaging method based on GNSS-R SAR, the present invention also proposes a corresponding fast range super-resolution imaging system based on GNSS-R SAR, which adopts the following technical solutions:

[0016] The fast range super-resolution imaging system based on GNSS-R SAR includes the following modules:

[0017] A preliminary imaging module is used to obtain GNSS echo signals, perform preliminary imaging of the GNSS echo signals using a range-Doppler algorithm, and express the preliminary imaging results in the form of a signal convolution model;

[0018] The observation matrix in the convolution model is the circulant matrix of the normalized autocorrelation function of the GNSS ranging code;

[0019] The measurement matrix reconstruction module is used to reduce the dimension of the measurement matrix using singular value decomposition. At the same time, a weight function is constructed and a weighted singular value preservation strategy is proposed to perform weighted reconstruction on the measurement matrix, thereby constructing a new signal convolution model.

[0020] The sparse prior regularization modeling module is used to construct the objective function for the new signal convolution model based on regularization, taking advantage of the sparse characteristics of the target and using the L1 norm as the constraint term, thus transforming the imaging problem into the inverse problem of sparse reconstruction.

[0021] The iterative solution module is used to design an accelerated iterative optimization algorithm to solve the objective function. The iterative optimization algorithm performs gradient descent and combines soft threshold shrinkage operation in each iteration, while introducing momentum terms to accelerate convergence.

[0022] And the output module outputs the fast range super-resolution imaging results based on GNSS-R SAR after the objective function is solved.

[0023] Furthermore, based on the GNSS-R SAR-based fast range super-resolution imaging method, the present invention also provides a computer device comprising a memory and one or more processors. The memory stores executable code. When the processor executes the executable code, the computer device implements the GNSS-R SAR-based fast range super-resolution imaging method.

[0024] In addition, based on the fast range super-resolution imaging method based on GNSS-R SAR, the present invention also proposes a computer-readable storage medium on which a program is stored.

[0025] When the program is executed by a processor, it is used to implement the steps of a fast range super-resolution imaging method based on GNSS-R SAR.

[0026] The present invention has the following advantages:

[0027] As described above, the present invention proposes a fast range super-resolution imaging method based on GNSS-R SAR for GNSS echo signals. The method of the present invention innovatively combines matrix compression, regularization modeling and fast optimization algorithms to form a set of efficient and stable range super-resolution processing procedures, and uses fast range super-resolution imaging technology to perform fast high-resolution imaging on GNSS echo data. Specifically, after obtaining the preliminary imaging results of the GNSS echo signal and giving its signal convolution model form, the observation matrix in the signal convolution model is first reduced in dimension by singular value decomposition. Based on the processing results of the weighted singular value preservation strategy, the signal convolution model is reconstructed using the inverse matrix of the truncated measurement matrix; then, starting from the regularization strategy, the L1 norm constraint is introduced to construct the objective function using the sparse characteristics of the target; finally, an accelerated iterative optimization algorithm is used to solve the objective function. Experiments have shown that the optimal range resolution of the original imaging algorithm is 150m. After processing by the method of the present invention, the range resolution can be increased to 60m, thereby significantly improving the range resolution of the GNSS echo data and effectively realizing the ability to distinguish close-range targets. The method of the present invention provides technical support for the practical application of GNSS-R SAR. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 Flowchart of a fast range super-resolution imaging method based on GNSS-R SAR in an embodiment of the present invention;

[0029] Figure 2 Schematic diagram of the GNSS-R SAR system configuration;

[0030] Figure 3 This is a flow chart of a fast optimization algorithm for the gradient search and threshold shrinkage mechanism in an embodiment of the present invention. DETAILED DESCRIPTION

[0031] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments:

[0032] Example 1

[0033] This embodiment 1 describes a fast range super-resolution imaging method based on GNSS-R SAR. This method combines matrix compression, regularized modeling and a fast optimization algorithm to form an efficient and stable range super-resolution processing flow, and then uses fast range super-resolution imaging technology to perform fast high-resolution imaging on GNSS echo data.

[0034] like Figure 1 As shown in FIG, the fast range super-resolution imaging method based on GNSS-R SAR includes the following steps:

[0035] Step 1. Obtain the GNSS echo signal and perform preliminary imaging of the GNSS echo signal using the range-Doppler algorithm. The preliminary imaging result is expressed as a signal convolution model.

[0036] GNSS-SAR system configuration is as follows Figure 2 As shown, its signal The form is:

[0037] .

[0038] in is the ranging code, is the navigation information code, then the two-dimensional forms of the direct signal and the echo signal are:

[0039] ;

[0040] .

[0041] in Indicates direct signal, Indicates the echo signal. For quick time, For slow time, represents the time delay of the direct signal, represents the time delay of the echo signal, 、 are the Doppler shifts of the direct signal and the echo signal, 、 are the phases of the direct signal and the navigation signal, respectively.

[0042] Figure 2 The middle reference channel Rc receives the direct signal from the satellite Tx as a reference signal, and the monitoring channel Sc receives the echo signal from the target Tg. Rt is the distance from Tx to Tg, Rr is the distance from Tg to Rx, and Rb is the distance from Tx to Rx.

[0043] First, a software receiver is used to process the data and generate a noise-free copy of the direct signal as a local reference signal. The reference signal is then cross-correlated with the echo signal in the fast time dimension to achieve range compression. The range compression result is derived as follows:

[0044] .

[0045] in is the result of distance compression, h is the cross-correlation function envelope of the ranging code, is the difference between the echo signal path and the direct signal path, is the center carrier frequency, c is the speed of light, Indicates; t is fast time, u is slow time.

[0046] Then, the range compression result is Fourier transformed in azimuth direction at slow time, and the results are as follows:

[0047] .

[0048] in represents the Fourier transform result, is the bistatic Doppler frequency, represents the slow time frequency, represents the slow time frequency shift, This completes the preliminary imaging of the GNSS echo signal.

[0049] The preliminary imaging results can be expressed as the following signal convolution model: .

[0050] in, is the echo signal after imaging, is the target scattering coefficient, is the noise, H is the measurement matrix, and the measurement matrix H is the circulant matrix of the normalized autocorrelation function of the GNSS ranging code, which can be expressed as follows:

[0051] .

[0052] Where 𝐿 represents the number of range sampling points during the PRI period. The problem of improving range resolution is transformed into solving a system of linear equations, a typical inverse problem. PRI is the duration of the C / A code, i.e., 1 ms.

[0053] Among them, the observation matrix H 、 、……、 、 Indicates the value of the GNSS ranging code autocorrelation function in 1ms.

[0054] This step 1 is GNSS echo signal reception and imaging pre-processing. By using a software receiver to obtain GNSS echo signals and performing preliminary imaging through a range-Doppler algorithm, low-resolution image data can be obtained.

[0055] Step 2. Use singular value decomposition to reduce the dimension of the observation matrix. At the same time, construct a weight function and propose a weighted singular value preservation strategy to reconstruct the observation matrix in a weighted manner, thereby constructing a new signal convolution model.

[0056] In traditional truncated singular value decomposition (SVD) methods, a truncation threshold w is often set to reduce computational complexity and mitigate the noise amplification effect caused by small singular values, retaining only the principal components corresponding to the first w larger singular values. However, this "hard truncation" strategy often ignores some small and medium-sized singular values ​​that carry useful information, thus affecting the recovery of image details.

[0057] In order to further improve the effect of super-resolution imaging, the present invention innovatively proposes a weighted singular value preservation strategy. The core idea of ​​this weighted singular value preservation strategy is not to completely discard small singular values ​​on the basis of truncated singular value decomposition, but to assign different weight coefficients to them according to the size of the singular values, thereby achieving more detailed retention of signal information.

[0058] Compared with the traditional hard truncation processing method, the weighted singular value retention strategy has the following advantages:

[0059] 1. Enhanced Information Preservation: Medium-magnitude singular values ​​are retained, helping to recover weakly scattered targets and low-intensity details. 2. Improved Image Quality: Compared to hard truncation, image details are more complete and edges are preserved more clearly. 3. Enhanced Noise Suppression: Noise components corresponding to small singular values ​​are given lower weight, avoiding direct noise amplification.

[0060] Specifically, the processing process of step 2 is as follows:

[0061] Step 2.1. Perform singular value decomposition (SVD) on the observation matrix H to obtain: .

[0062] in and is a matrix with orthogonal columns, and Representation matrix and The elements in the matrix , is a singular value, satisfying .

[0063] L represents the number of range sampling points during the PRI period (ie, the duration of the C / A code, ie, 1 ms).

[0064] Step 2.2. Construct the weight function.

[0065] Design a monotonically decreasing weight function , the calculation formula is as follows:

[0066] .

[0067] in is a tuning parameter used to control the weight decay rate. is the largest singular value, .

[0068] Step 2.3. Based on the weight function, a weighted singular value preservation strategy is proposed to reconstruct the observation matrix in a weighted manner and construct a new weighted reconstruction matrix. ; The calculation formula of the weighted singular value retention strategy is as follows:

[0069] .

[0070] Then the echo signal y after imaging in the signal convolution model is reconstructed accordingly as , n is reconstructed as , the calculation formula is as follows:

[0071] ; .

[0072] in, Represents the noise, and then obtains the new signal convolution model. The calculation formula is as follows: .

[0073] This step 2 is the construction of the ranging code correlation matrix and the matrix reconstruction process of the weighted singular value retention strategy. The autocorrelation function of the GNSS ranging code is constructed as an observation matrix. To avoid the computational complexity caused by the high dimension of the matrix, a weighted singular value retention strategy is proposed to address the problem of "hard truncation" in traditional truncated singular value decomposition, which easily loses weak target information. While compressing the matrix dimension, it retains the information contained in some small and medium singular values, taking into account both computational efficiency and imaging accuracy.

[0074] Step 3. For the new signal convolution model, starting from regularization, taking advantage of the sparse characteristics of the target, using the L1 norm as a constraint term to construct the objective function, the imaging problem is transformed into an inverse problem of sparse reconstruction.

[0075] Starting from regularization, taking advantage of the sparse characteristics of the target and taking the L1 norm as the constraint term, the constructed objective function is as follows:

[0076] .

[0077] in is the estimated value obtained by the optimization algorithm, x is the target scattering coefficient to be solved, is the regularization parameter.

[0078] Step 3 is the sparse prior regularization modeling process. Taking into account the sparsity characteristics of the target scene, the L1 norm regularization term is introduced to construct the objective function, transforming the imaging problem into the inverse problem of sparse reconstruction. Next, to efficiently solve this sparse reconstruction problem, the present invention proposes a fast optimization algorithm that combines gradient search and threshold shrinkage mechanisms.

[0079] Step 4. Design an accelerated iterative optimization algorithm to solve the objective function. The iterative optimization algorithm performs gradient descent combined with soft threshold shrinkage operation in each iteration, and introduces momentum term to speed up convergence.

[0080] like Figure 3 As shown in Figure 2, the process of solving the objective function using the accelerated iterative optimization algorithm is as follows:

[0081] Step 4.1. Initialize the image estimate.

[0082] Initial estimate , momentum auxiliary variable , momentum parameter ; Set the number of loop iterations k and the maximum number of iterations k max , and let the initial value of the number of loop iterations k be 1.

[0083] Step 4.2. Calculate the gradient direction and update the estimate.

[0084] At the kth iteration, the current estimate is , calculate the gradient, the calculation formula is as follows:

[0085] .

[0086] It is the gradient direction of the data residual to the image, which indicates the sensitivity of the error in the current imaging estimation to the target.

[0087] Do a step of gradient descent update, the calculation formula is as follows:

[0088] .

[0089] in represents the gradient, is the adaptive step size, which is used to control the scale of the search direction.

[0090] Step 4.3. Apply sparsity soft thresholding.

[0091] In order to suppress noise and emphasize the sparse structure of the image, a soft threshold function is introduced , used for sparsity constraint, suppresses the coefficients with amplitudes less than the threshold to 0 and weakens the amplitudes of the rest, thereby retaining the main reflection target in the image and removing background noise. This mechanism helps to improve the sparsity and clarity of the imaging results. Its calculation formula is as follows:

[0092] .

[0093] in, represents the image estimation value of the kth iteration, represents the symbolic function, Represents the maximum function.

[0094] This step 4.3 introduces a soft threshold function to compress components with smaller absolute values ​​to zero while retaining scattered components greater than the threshold, thereby enhancing the boundary and point features of the image, removing most of the noise information, and improving the signal-to-noise ratio of the target.

[0095] Step 4.4. Introduce historical information to accelerate convergence.

[0096] In order to avoid the oscillation problem caused by pure gradient descent, the momentum update strategy is adopted. First, the momentum parameter is updated. The formula is as follows:

[0097] .

[0098] in, represents the momentum of the previous iteration, Represents the momentum of the current iteration.

[0099] Then update the search point for the next iteration, which is the historical guide. The calculation formula is as follows:

[0100] .

[0101] in represents the estimated value at the k+1th iteration, Represents the historical information at the k-1th iteration.

[0102] The first formula in step 4.4 is to update the momentum, and the second formula is to construct the next search point based on historical information. This momentum update strategy essentially utilizes the "inertia" information in the gradient direction, which can effectively speed up the iteration speed, improve the convergence efficiency, and suppress the oscillation behavior that may occur in simple gradient descent. This momentum update mechanism is equivalent to making a weighted prediction between the current solution and the previous solution, thereby utilizing the "inertia direction" to accelerate the convergence speed and reduce oscillation.

[0103] Step 4.5. Repeat the iterative process from step 4.2 to step 4.4 above until the maximum number of iterations is reached or the estimated value obtained by the iterative optimization algorithm is reached. When the error threshold is met, that is, the decrease in the objective function value is less than the threshold, the iteration ends.

[0104] The estimated value obtained after the objective function is solved , which is the result of fast distance super-resolution imaging.

[0105] In step 4, the objective function is solved by adopting a new accelerated iterative optimization algorithm. This optimization algorithm performs gradient descent and combines it with a soft threshold shrinkage operation in each iteration. At the same time, a momentum term is introduced to accelerate convergence. This optimization algorithm takes into account both accuracy and computational speed and is particularly suitable for high-resolution image reconstruction in GNSS low signal-to-noise ratio environments.

[0106] Step 5. After the objective function is solved, the fast range super-resolution imaging results based on GNSS-R SAR are output.

[0107] After the above-mentioned fast distance super-resolution processing, the resolution of the output imaging results is significantly improved. The final output imaging results can effectively resolve targets at a distance of 60 meters, which is a significant improvement compared to the original 150-meter distance resolution.

[0108] Compared with traditional methods, this invention innovatively combines matrix compression, regularized modeling, and fast optimization algorithms to form an efficient and stable range super-resolution processing flow. Furthermore, this invention introduces the concept of a weighted singular value retention strategy into the dimensionality reduction process of the GNSS ranging code construction matrix for the first time, and designs an adaptive weight function to optimize the signal-to-noise distinction. Combined with a sparse reconstruction framework, this method forms an efficient GNSS-R SAR range super-resolution imaging algorithm. In terms of range resolution, it is improved from 150 meters in traditional imaging algorithms to 60 meters, effectively achieving the ability to distinguish close-range targets.

[0109] Example 2

[0110] This embodiment 2 describes a fast range super-resolution imaging system based on GNSS-R SAR. This system is based on the same inventive concept as the fast range super-resolution imaging method based on GNSS-R SAR described in the above embodiment 1.

[0111] A fast range super-resolution imaging system based on GNSS-R SAR, including the following modules:

[0112] A preliminary imaging module is used to obtain GNSS echo signals, perform preliminary imaging of the GNSS echo signals using a range-Doppler algorithm, and express the preliminary imaging results in the form of a signal convolution model;

[0113] The observation matrix in the convolution model is the circulant matrix of the normalized autocorrelation function of the GNSS ranging code;

[0114] The measurement matrix reconstruction module is used to reduce the dimension of the measurement matrix using singular value decomposition. At the same time, a weight function is constructed and a weighted singular value preservation strategy is proposed to perform weighted reconstruction on the measurement matrix, thereby constructing a new signal convolution model.

[0115] The sparse prior regularization modeling module is used to construct the objective function for the new signal convolution model based on regularization, taking advantage of the sparse characteristics of the target and using the L1 norm as the constraint term, thus transforming the imaging problem into the inverse problem of sparse reconstruction.

[0116] The iterative solution module is used to design an accelerated iterative optimization algorithm to solve the objective function. The iterative optimization algorithm performs gradient descent and combines soft threshold shrinkage operation in each iteration, while introducing momentum terms to accelerate convergence.

[0117] And the output module outputs the fast range super-resolution imaging results based on GNSS-R SAR after the objective function is solved.

[0118] It should be noted that, in the fast range super-resolution imaging system based on GNSS-R SAR in this embodiment 2, the implementation process of the functions and effects of each functional module is detailed in the implementation process of the corresponding steps of the method in the above embodiment 1, and will not be repeated here.

[0119] Example 3

[0120] This embodiment 3 describes a computer device. The computer device includes a memory and one or more processors. The memory stores executable code. When the processor executes the executable code, it is used to implement the steps of the GNSS-R SAR-based fast range super-resolution imaging method described in the above embodiment 1.

[0121] In this embodiment, the computer device is any device or apparatus with data processing capability, which will not be described in detail here.

[0122] Example 4

[0123] This embodiment 4 describes a computer-readable storage medium having a program stored thereon. When the program is executed by a processor, it is used to implement the steps of the fast range super-resolution imaging method based on GNSS-R SAR in the above embodiment 1.

[0124] The computer-readable storage medium can be an internal storage unit of any device or apparatus with data processing capabilities, such as a hard disk or memory, or an external storage device of any device with data processing capabilities, such as a plug-in hard disk, smart media card (SMC), SD card, flash card, etc. equipped on the device.

[0125] Of course, the above description is only a preferred embodiment of the present invention, and the present invention is not limited to the above-mentioned embodiments. It should be noted that all equivalent substitutions and obvious deformation forms made by any technician familiar with this field under the guidance of this specification fall within the substantive scope of this specification and should be protected by the present invention.

Claims

1. A fast range super-resolution imaging method based on GNSS-R SAR, characterized by: The steps include: Step 1. Obtain GNSS echo signals and perform preliminary imaging of the GNSS echo signals using the range-Doppler algorithm. The preliminary imaging results are expressed as a signal convolution model. The observation matrix in the signal convolution model is the circulant matrix of the normalized autocorrelation function of the GNSS ranging code; Step 2. Use singular value decomposition to reduce the dimension of the measurement matrix. At the same time, construct a weight function and propose a weighted singular value preservation strategy to reconstruct the measurement matrix in a weighted manner, thereby constructing a new signal convolution model. Step 3. Based on the new signal convolution model, starting from regularization, taking advantage of the sparse nature of the target, and using the L1 norm as a constraint term to construct the objective function, the imaging problem is transformed into an inverse problem of sparse reconstruction. Step 4. Design an accelerated iterative optimization algorithm to solve the objective function. The iterative optimization algorithm performs gradient descent combined with soft threshold shrinkage in each iteration, and introduces a momentum term to accelerate convergence. Step 5. After the objective function is solved, the fast range super-resolution imaging results based on GNSS-R SAR are output.

2. The fast range super-resolution imaging method based on GNSS-R SAR according to claim 1, characterized in that: In step 1, the preliminary imaging process of the GNSS echo signal is as follows: First, a software receiver is used to process the data and generate a noise-free copy of the direct signal as a local reference signal. The reference signal is then cross-correlated with the echo signal in the fast time dimension to achieve range compression. The range compression result is derived as follows: ； in is the result of distance compression, h is the cross-correlation function envelope of the ranging code, is the difference between the echo signal path and the direct signal path, is the center carrier frequency, c is the speed of light, is noise, t is fast time, and u is slow time; Then, the range compression result is Fourier transformed in azimuth direction at slow time, and the results are as follows: ； in represents the Fourier transform result, is the bistatic Doppler frequency, represents the slow time frequency, represents the slow-time frequency shift, Indicates noise; This completes the preliminary imaging of the GNSS echo signal.

3. The fast range super-resolution imaging method based on GNSS-R SAR according to claim 1, characterized in that: In step 1, the preliminary imaging results are expressed as a signal convolution model in the form of: ;in, is the echo signal after imaging, is the target scattering coefficient, is noise, is the measurement matrix, the measurement matrix is the circulant matrix of the normalized autocorrelation function of the GNSS ranging code.

4. The fast range super-resolution imaging method based on GNSS-R SAR according to claim 3, characterized in that: The step 2 is specifically as follows: Step 2.

1. Perform singular value decomposition on the observation matrix H to obtain: ; in and is a matrix with orthogonal columns, and Representation matrix and The elements in the matrix , is a singular value, satisfying ; L represents the number of range sampling points during PRI, where PRI is the duration of the C / A code; Step 2.

2. Construct the weight function: Design a monotonically decreasing weight function , the calculation formula is as follows: ； in is a tuning parameter used to control the weight decay rate. is the largest singular value, ; Step 2.

3. Based on the weight function, a weighted singular value preservation strategy is proposed to reconstruct the observation matrix in a weighted manner and construct a new weighted reconstruction matrix. ; The calculation formula of the weighted singular value retention strategy is as follows: ； Then the echo signal y after imaging in the signal convolution model is reconstructed accordingly as , n is reconstructed as , the formula is as follows: ， ； in, Represents the noise, and then obtains the new signal convolution model. The calculation formula is as follows: .

5. The fast range super-resolution imaging method based on GNSS-R SAR according to claim 4, characterized in that: In step 3, the objective function formula constructed is as follows: ； Where x is the target scattering coefficient to be solved, is the estimated value obtained by the iterative optimization algorithm, is the regularization parameter.

6. The fast range super-resolution imaging method based on GNSS-R SAR according to claim 5, characterized in that: In step 4, the process of solving the objective function using the accelerated iterative optimization algorithm is as follows: Step 4.

1. Initialize the image estimate; Initial estimate , momentum auxiliary variable , momentum parameter ; Set the number of loop iterations k and the maximum number of iterations k max , and let the initial value of the loop iteration number k be 1; Step 4.

2. Calculate the gradient direction and update the estimated value; At the kth iteration, the current estimate is , calculate the gradient, the calculation formula is as follows: ； It is the gradient direction of the data residual to the image, which indicates the sensitivity of the error in the current imaging estimation to the target; Do a step of gradient descent update, the calculation formula is as follows: ； in represents the gradient, is the adaptive step size, used to control the scale of the search direction; Step 4.

3. Apply sparsity soft thresholding. In order to suppress noise and emphasize the sparse structure of the image, a soft threshold function is introduced , the calculation formula is as follows: ； in, represents the image estimation value of the kth iteration, represents the symbolic function, represents the maximum value function; Step 4.

4. Introduce historical information to accelerate convergence; In order to avoid the oscillation problem caused by pure gradient descent, the momentum update strategy is adopted. First, the momentum parameter is updated. The formula is as follows: ； in, represents the momentum of the previous iteration, Represents the momentum of the current iteration; Then update the search point for the next iteration, and the calculation formula is as follows: ； in represents the estimated value at the k+1th iteration, Represents the historical information at the k-1th iteration; Step 4.

5. Repeat the iterative process from step 4.2 to step 4.4 above until the maximum number of iterations is reached or the estimated value obtained by the iterative optimization algorithm is reached. If the error threshold is met, the iteration ends; The estimated value obtained after the objective function is solved , which is the result of fast distance super-resolution imaging.

7. A fast range super-resolution imaging system based on GNSS-R SAR, characterized by: Includes the following modules: A preliminary imaging module is used to obtain GNSS echo signals, perform preliminary imaging of the GNSS echo signals using a range-Doppler algorithm, and express the preliminary imaging results in the form of a signal convolution model; The observation matrix in the signal convolution model is the circulant matrix of the normalized autocorrelation function of the GNSS ranging code; The measurement matrix reconstruction module is used to reduce the dimension of the measurement matrix using singular value decomposition. At the same time, a weight function is constructed and a weighted singular value preservation strategy is proposed to perform weighted reconstruction on the measurement matrix, thereby constructing a new signal convolution model. The sparse prior regularization modeling module is used to construct the objective function for the new signal convolution model based on regularization, taking advantage of the sparse characteristics of the target and using the L1 norm as the constraint term, thus transforming the imaging problem into the inverse problem of sparse reconstruction. The iterative solution module is used to design an accelerated iterative optimization algorithm to solve the objective function. The iterative optimization algorithm performs gradient descent and combines soft threshold shrinkage operation in each iteration, while introducing momentum terms to accelerate convergence. And the output module outputs the fast range super-resolution imaging results based on GNSS-R SAR after the objective function is solved.

8. A computer device comprising a memory and one or more processors; executable code is stored in the memory; and When the processor executes the executable code, it is used to implement the steps of the GNSS-R SAR based fast range super-resolution imaging method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a program stored thereon; characterized in that: When the program is executed by a processor, it is used to implement the steps of the GNSS-R SAR based fast range super-resolution imaging method according to any one of claims 1 to 6.

Images