A joint optimization SAR imaging method combining a learnable transmit pattern and an imaging algorithm

Abstract

The application provides a joint optimization SAR imaging method combining a learnable emission mode and an imaging algorithm, and relates to the field of synthetic aperture radar (SAR) imaging. The method comprises the following steps: regarding a random part in a non-uniform pulse emission time as a learnable variable, simulating a non-uniform echo signal under a current emission mode, obtaining a SAR image through a deep unfolding SAR imaging algorithm based on an approximate observation and a neural network sparse domain mapping, and taking the SAR image as an optimization target. Meanwhile, parameters in the pulse emission mode and the imaging algorithm are optimized, so that a non-uniform pulse emission mode sequence with optimal anti-aliasing performance and a matching SAR imaging algorithm are obtained. The application combines the learnable pulse emission mode and the SAR imaging algorithm, and takes the final SAR imaging result as the optimization target, so that the imaging algorithm fully utilizes the advantages of the non-uniform emission mode, and the quality of imaging is improved as much as possible while the range is extended.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of synthetic aperture radar imaging, and in particular to a joint optimization SAR imaging method combining a learnable transmit pattern and an imaging algorithm. BACKGROUND

[0002] Most of the existing synthetic aperture radar (SAR) imaging algorithms based on compressed sensing use random sampling to obtain down-sampled echo signals, which only reduces the amount of data and does not change the minimum interval of adjacent pulses, so as to fail to expand the mapping bandwidth. The SAR imaging algorithms for realizing wide mapping band often decouple the transmit pattern and the imaging algorithm for separate design, which causes the imaging algorithm to fail to fully utilize the anti-aliasing advantage of the non-uniform transmit pattern, thereby affecting the imaging quality. SUMMARY

[0003] The present application aims to at least partially solve one of the problems in the related art.

[0004] To this end, a first object of the present application is to propose a joint optimization SAR imaging method combining a learnable transmit pattern and an imaging algorithm, which combines a learnable pulse transmit pattern and a SAR imaging algorithm, takes the final SAR imaging result as an optimization target, and makes the imaging algorithm fully utilize the advantage of the non-uniform transmit pattern to expand the range while improving the imaging quality as much as possible.

[0005] A second object of the present application is to propose an electronic device.

[0006] A third object of the present application is to propose a computer-readable storage medium.

[0007] A fourth object of the present application is to propose a computer program product.

[0008] To achieve the above objects, a first aspect of the present application proposes a joint optimization SAR imaging method combining a learnable transmit pattern and an imaging algorithm, comprising:

[0009] S1. Generating a non-uniform pseudo echo signal according to SAR system parameters and a simulation scene as an input signal of an imaging iteration process, and initializing an imaging result matrix as a zero matrix as an initial state of the imaging iteration process;

[0010] S2. Based on the current imaging result, first generating an approximate echo by an inverse imaging operator, calculating the residual error between the approximate echo and the simulation echo, then obtaining the imaging result of the residual error by an imaging operator, and performing linear update on the imaging result of the last round by using the imaging result of the residual error;

[0011] S3, mapping the updated linear imaging result to the sparse domain and filtering out noise using a soft threshold iteration function to obtain a nonlinear imaging result;

[0012] S4, repeating the linear and nonlinear updating steps until a set number of iterations is reached to obtain a final imaging result;

[0013] S5, after the iteration is completed, calculating the loss function and the gradient, and using a random parameter freezing strategy for the pulse emission time to randomly select a portion of the parameter gradient freezing at the pulse emission time;

[0014] S6, updating the pulse emission mode and the imaging algorithm parameters according to the updated gradient of the random parameter freezing strategy and performing back propagation and updating;

[0015] S7, continuously repeating steps S2-S6 until convergence is achieved to obtain the final optimized pulse emission mode and matching imaging algorithm.

[0016] Optionally, the non-uniform pseudo echo signal is generated according to the SAR system parameters and the simulation scene as the input signal of the imaging iteration process, and the imaging result matrix is initialized as a zero matrix as the initial state of the imaging iteration process, comprising:

[0017] According to the SAR system parameters Θ and the simulation scene X, a simulated azimuth non-uniform echo signal Y is generated as the input signal of the imaging iteration process;

[0018] The imaging result X 0 is initialized as a zero matrix as the initial state of the imaging iteration process.

[0019] Optionally, the residual error between the approximate echo and the simulation echo is calculated based on the current imaging result, and the imaging result is linearly updated using the residual error, comprising:

[0020] Based on the imaging result X k-1 of the previous round, an approximate echo matrix G(X k-1 ) is obtained according to the approximate observation inverse imaging operator, wherein G() represents the inverse imaging operator;

[0021] The residual error between the approximate echo matrix G(X k-1 ) and the simulation echo Y is calculated, and the approximate imaging result H(Y-G(X k-1 )) of the residual error is obtained according to the imaging operator, wherein H() represents the imaging operator;

[0022] The product of the residual error coefficient and the approximate imaging result H(Y-G(X k-1 )) is added to the imaging result X k-1 of the previous round to obtain the updated linear imaging result, and the expression is:

[0023] Z k =X k-1 +α k ·H(YG(X k-1 ))

[0024] Among them, Z k This represents the linear imaging result after the k-th iteration update.

[0025] Optionally, the step of mapping the updated linear imaging result to a sparse domain and filtering out noise using a soft-thresholding iterative function to obtain a nonlinear imaging result includes:

[0026] Separate the linear imaging result Z as a complex matrix k The real and imaginary parts are used to transform it into a real matrix D(Z). k );

[0027] The real matrix D(Z) is processed by three layers of convolution Conv. k The result is obtained by mapping the data to a sparse domain and using a soft-thresholding iterative function to filter out noise. Among them, F k This represents the processing procedure of a three-layer convolutional Conv. This represents the soft threshold iteration function;

[0028] The noise-filtered result is then back-mapped back to the image domain using a three-layer convolution Conv, and used as the residual along with the initial complex matrix Z. k To add to the result, we obtain the updated nonlinear imaging result, expressed as:

[0029]

[0030] in, This represents the process of mapping the result of a three-layer convolution Conv back to the image domain, where C is the constant coefficient for residual adjustment, and X... k This represents the nonlinear imaging result after the k-th iteration update.

[0031] Optionally, the calculation of the loss function includes:

[0032] Calculate the less-than-zero penalty term for the pulse emission mode.

[0033] Penalty term in computational imaging algorithms for mapping between the sparse domain and the image domain

[0034] Calculate the mean square error (MSE) difference between the final imaging function and the target image.

[0035] The loss function is calculated as follows:

[0036]

[0037] wherein, L (Γ Total ) is a loss value of the loss function, γ1, γ2, γ3 are weight coefficients of different loss terms in the loss function, is the total number of pulses in the pulse emission mode, is the emission parameter of the n a th pulse, J is the total number of imaging images, j is the number of imaging images, K is the set number of iterations, GT is the target image, X K is the final imaging result, denotes the L2 norm.

[0038] Optionally, the random parameter freezing strategy is used for the pulse emission time, and a part of the parameter gradient freezing is randomly selected in the pulse emission time, including:

[0039] The random parameter freezing strategy is applied, and a group of mask sequences with the same length as the pulse emission time is randomly generated, and the expression is:

[0040]

[0041] wherein, each element in the mask sequence is a random variable subject to Bernoulli distribution;

[0042] The mask sequence is multiplied by the gradient of the loss function with respect to the pulse emission mode d η , so as to achieve the purpose of random zeroing, and the expression is:

[0043]

[0044] wherein, denotes the gradient calculation of the loss function with respect to the pulse emission mode d η , to obtain the corresponding gradient vector; denotes the item-by-item multiplication of the mask sequence m and the gradient vector.

[0045] To achieve the above purpose, a second aspect embodiment of the present application provides an electronic device, comprising: a processor, and a memory in communication connection with the processor;

[0046] The memory stores computer execution instructions;

[0047] The processor executes the computer execution instructions stored in the memory, to realize the method as described in any one of the first aspect.

[0048] To achieve the above object, the third aspect of the present application provides a computer readable storage medium, wherein the computer readable storage medium stores computer execution instructions, and the computer execution instructions are executed by a processor to implement the method according to any one of the first aspect.

[0049] To achieve the above object, the fourth aspect of the present application provides a computer program product, wherein the computer program is executed by a processor to implement the method according to any one of the first aspect.

[0050] The technical scheme provided by the embodiments of the present application at least brings the following beneficial effects:

[0051] By combining the learnable pulse emission mode with the SAR imaging algorithm, the imaging algorithm can fully utilize the anti-aliasing advantage of the non-uniform emission mode, and successfully realizes the expansion of the wide swath. The depth unfolding method based on approximate observation is adopted to significantly reduce the computational complexity, the random parameter freezing strategy is introduced to improve the randomness and robustness of the pulse emission mode, thereby effectively improving the resolution and quality of the SAR imaging, and meeting the imaging requirements of the high-resolution wide swath.

[0052] Additional aspects and advantages of the present application will be in part apparent and in part pointed out hereinafter. BRIEF DESCRIPTION OF DRAWINGS

[0053] The above and / or additional aspects and advantages of the present application will become apparent and be readily appreciated from the following description, taken in conjunction with the accompanying drawings, in which:

[0054] Figure 1 A flowchart of a joint optimization SAR imaging method combining a learnable emission mode and an imaging algorithm provided by an embodiment of the present application;

[0055] Figure 2 A flowchart of a joint optimization SAR imaging method combining a learnable emission mode and an imaging algorithm provided by an embodiment of the present application;

[0056] Figure 3 A flowchart of a joint optimization SAR imaging method combining a learnable emission mode and an imaging algorithm provided by an embodiment of the present application. DETAILED DESCRIPTION

[0057] The embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference signs represent the same or similar elements or elements having the same or similar functions throughout. The embodiments described below are exemplary and are intended to explain the present application, and cannot be understood as limiting the present application.

[0058] Most of the existing SAR imaging algorithms based on compressed sensing use random sampling to obtain down-sampled echo signals, which only reduces the amount of data and does not change the minimum interval of adjacent pulses, so it is impossible to expand the mapping bandwidth. The SAR imaging algorithm for realizing wide mapping band often decouples the transmission mode and the imaging algorithm for separate design, which causes the imaging algorithm to be unable to fully utilize the anti-aliasing advantage of the non-uniform transmission mode, thereby affecting the imaging quality.

[0059] In order to overcome the problem that the existing wide SAR imaging algorithm cannot fully remove the aliasing caused by the non-uniform transmission and thus affects the imaging quality, the embodiments of the present application provide a joint optimization SAR imaging method combining a learnable transmission mode and an imaging algorithm, Figure 1 、 2 and 3 are flowcharts of a joint optimization SAR imaging method combining a learnable transmission mode and an imaging algorithm provided by the embodiments of the present application.

[0060] It should be noted that before the flow starts, the following input parameters need to be set:

[0061] SAR simulation scene X: used to simulate the real scene of SAR imaging, providing the expected imaging target for the algorithm.

[0062] SAR system parameters Θ: important parameters for controlling the characteristics of the SAR system, including transmission frequency, bandwidth, sampling rate, etc., which directly affect the characteristics of the generated echo signal.

[0063] Iteration number K: controls the number of updates of the imaging algorithm, used to adjust the accuracy of the optimization process (set to 4 times in the present application).

[0064] Loss function coefficients γ1, γ2, γ3: respectively control the weights of different parts in the loss function, used to balance the imaging quality, anti-aliasing performance and restrictions on the transmission mode.

[0065] As shown in Figure 1 , the method comprises the following steps:

[0066] Step S1, according to the SAR system parameters and the simulation scene, a non-uniform pseudo echo signal is generated as the input signal of the imaging iteration process, and the imaging result matrix is initialized as a zero matrix as the initial state of the imaging iteration process.

[0067] In the embodiments of the present application, first, according to the SAR system parameters Θ and the simulation scene X, a simulated azimuth non-uniform echo signal Y is generated as the input signal of the imaging iteration process.

[0068] It can be understood that the generated simulation signal Y is the input signal of the entire imaging process, which simulates the real SAR data and provides a basic reference for the subsequent imaging algorithm. The generated simulation signal not only provides actual observation data for the subsequent steps, but also defines the imaging bandwidth requirements and signal distribution of the system, providing an optimization target for the imaging algorithms of steps S2 and S3.

[0069] In addition, the imaging result X 0 is initialized to a zero matrix as the initial state of the imaging iteration process.

[0070] It can be understood that the zero matrix initialization provides an unbiased starting point, so that the imaging process starts from a consistent basic state. This initialization step ensures that the subsequent iterative updates can be optimized without initial bias, thereby ensuring the stability of the imaging algorithm.

[0071] In addition, the initial imaging result matrix will be directly passed as input to the linear update part in step S2, starting the step-by-step optimization of the imaging.

[0072] In step S2, based on the current imaging result, the approximate echo is first generated by the inverse imaging operator, the residual between the approximate echo and the simulation echo is calculated, and the imaging result of the residual is obtained by using the imaging operator, and the imaging result of the residual is used to linearly update the imaging result of the last round.

[0073] In the embodiments of the present application, in each iteration, linear update is used to gradually approach the target imaging result.

[0074] Specifically, in each update, the imaging result X k-1 of the last round is inversely projected by using the inverse imaging operator G() to obtain the approximate echo matrix G(X k-1 ), and the inverse imaging operator is used to simulate the process of returning from the image domain to the frequency domain in the imaging process, in order to generate the observation signal corresponding to the current imaging result.

[0075] Then, the difference between the approximate echo matrix G(X k-1 ) and the simulation echo Y is calculated to generate a residual matrix, and the residual matrix represents the deviation between the current imaging result and the real observation signal.

[0076] Subsequently, the residual matrix is projected according to the imaging operator to obtain the correction information mapped from the observation domain back to the image domain, that is, the approximate imaging result H(Y-G(X k-1 )), wherein H() represents the imaging operator. The imaging operator projects the frequency domain residual information back to the image domain, thereby applying a correction to the image representation of the current imaging result

[0077] Finally, the approximate imaging result H(Y-G(Xk-1 )) is weighted and superimposed to the imaging result X of the previous round k-1 , to obtain the updated linear imaging result Z

[0078] Z k = X k-1 + a k · H(Y - G(X k-1 ))

[0079] wherein Z k is the updated linear imaging result of the kth iteration.

[0080] Unlike the conventional observation matrix method, the embodiments of the present application use the inverse imaging operator and the imaging operator in cooperation, so that the imaging process can fully utilize the information in the frequency domain and the image domain in each iteration, and can convert the multiplication operation of a large matrix into multiple multiplication operations of small matrices, greatly reducing the time complexity. This process not only ensures the effective correction between the observation signal and the imaging result, but also enables the algorithm to gradually approach the true imaging result under low computational complexity.

[0081] Step S3, mapping the updated linear imaging result to the sparse domain and using a soft threshold iteration function to filter out noise to obtain a nonlinear imaging result.

[0082] In the embodiments of the present application, the image can also be nonlinearly updated in each iteration to further refine the imaging result and improve the image details.

[0083] Specifically, first, separate the real part and the imaginary part of the linear imaging result Z k as a complex matrix, and convert it into a real number matrix D(Z k ).

[0084] Then, map the real number matrix D(Z k ) to the sparse domain through three-layer convolution Conv, remove noise and enhance the sparsity of the image.

[0085] Further, use a soft threshold iteration function to filter out noise and retain key image features to obtain a result after filtering out noise wherein F k represents the processing process of the three-layer convolution Conv, and the soft threshold iteration function.

[0086] Finally, map the result after filtering out noise back to the image domain through the three-layer convolution Conv again, and take it as the residual error and the initial complex matrix Z k to obtain the updated nonlinear imaging result, and the expression is:

[0087]

[0088] wherein, represents the process of three-layer convolution Conv reflecting the result back to the image domain, C is a constant coefficient of residual adjustment, X k is the nonlinear imaging result after the kth iteration update.

[0089] It should be noted that the nonlinear update effectively removes residual noise and enhances image details through sparse mapping and soft threshold iteration, so that the imaging result is further optimized in each iteration, and the introduction of sparse mapping expands the application of the algorithm in non-sparse scenarios, ensuring higher imaging quality.

[0090] Step S4, repeating the linear and nonlinear update steps until a set number of iterations is reached, to obtain the final imaging result.

[0091] In the embodiments of the present application, steps S2 and S3 are repeated, and linear and nonlinear updates are alternately performed until a preset number of iterations K is reached, and the iteration process is stopped to obtain the final imaging result X K .

[0092] Through the alternating optimization of linear and nonlinear updates, the imaging result gradually approaches the target image in each round, and the optimization result is continuously accumulated, finally realizing high-precision imaging. The iterative optimization process not only reduces the error, but also effectively processes noise and details, so that the imaging quality gradually improves.

[0093] Step S5, after the iteration is completed, the loss function and the gradient are calculated, and a random parameter freezing strategy is used for the pulse emission time, and a part of the parameter gradient is frozen at the pulse emission time.

[0094] After the iteration is completed, the error of the current imaging result is evaluated by calculating the loss function, and feedback is provided for subsequent optimization.

[0095] In the embodiments of the present application, the loss function mainly includes three parts, which are:

[0096] (1) A less than zero penalty term for the pulse emission mode Used to limit the parameter range in the emission mode to ensure reasonable parameters;

[0097] (2) A penalty term for the mapping of the imaging algorithm in the sparse domain and the image domain Used to constrain the error of the imaging algorithm in the grid domain and the image mapping process;

[0098] (3) The mean square error (MSE) gap between the final imaging function and the target image Measure the error between the final imaging result and the target image.

[0099] In the embodiments of the present application, the expression of the loss function is:

[0100]

[0101] wherein L (Γ Total ) is the loss value of the loss function, γ1, γ2, γ3 are weight coefficients of different loss terms in the loss function, is the total number of pulses in the pulse emission mode, is the emission parameter of the i na th pulse, J is the total number of imaging images, j is the number of imaging images, K is the set number of iterations, GT is the target image, X K is the final imaging result, represents the L2 norm.

[0102] It should be noted that the calculation result of the loss function provides quantitative feedback for the optimization process, guiding the adjustment direction of the pulse emission mode and the imaging parameter in the next iteration. By minimizing the loss function, the imaging algorithm can gradually adjust each parameter to make the imaging result approach the optimal solution, thereby improving the imaging accuracy and anti-aliasing ability.

[0103] In addition, the embodiments of the present application also introduce a random parameter freezing strategy in the back propagation process to enhance the randomness and robustness of the pulse emission mode.

[0104] Specifically, first, a random parameter freezing strategy is applied to randomly generate a group of mask sequences with the same length as the pulse emission time, and the expression is:

[0105]

[0106] wherein each element in the mask sequence is a random variable subject to Bernoulli distribution, that is, each element takes the value of 1 with a probability of 70%.

[0107] Then, the mask sequence is multiplied by the gradient of the loss function with respect to the pulse emission mode d η , and 30% of the elements in the gradient are randomly frozen to achieve the purpose of random zero, and the expression is:

[0108]

[0109] wherein represents gradient calculation of the loss function with respect to the pulse emission mode d η , to obtain the corresponding gradient vector; represents multiplying the mask sequence m with the gradient vector item by item.

[0110] It should be noted that the random part of the pulse emission mode is taken as a learnable variable, and the gradient of part of the sequence elements is set to zero in the updating process through the random parameter freezing strategy, which improves the anti-interference of the emission mode and makes the imaging algorithm have stronger robustness in complex environments. In addition, the freezing strategy prevents overfitting of the model during optimization, enabling the algorithm to maintain better generalization and stability.

[0111] Step S6, the updated gradient according to the random parameter freezing strategy is used to perform back propagation on the parameters of the pulse emission mode and the imaging algorithm and update them.

[0112] In the embodiments of the present application, the optimized gradient makes the pulse emission mode have higher randomness and robustness through the random parameter freezing strategy, reducing the performance degradation caused by overfitting. In addition, the strategy of freezing part of the gradient enhances the anti-aliasing capability, making the emission mode more adaptable in complex environments. At the same time, the parameters of the imaging algorithm are dynamically optimized through back propagation, further improving the resolution and quality of the imaging results.

[0113] Step S7, steps S2-S6 are repeatedly performed until convergence, obtaining the final optimized pulse emission mode and matching imaging algorithm.

[0114] In the embodiments of the present application, the optimization steps S2-S6 are repeatedly performed until convergence, finally obtaining the optimized pulse emission mode d η and the imaging algorithm parameters, achieving the goal of high resolution and wide swath imaging. In the wide swath scenario, the optimized pulse emission mode has stronger anti-aliasing capability and can achieve wide imaging with high resolution; the matching imaging algorithm significantly improves the imaging clarity and detail performance while maintaining high computational efficiency, meeting the high resolution imaging demand.

[0115] In actual SAR systems, the pulse emission and echo data collection can be performed according to the trained pulse emission time, and then the trained imaging algorithm is used to obtain the imaging result.

[0116] In terms of SAR imaging effect, in order to verify the superiority of the method of the present application, the method of the present application is compared with other literature methods, and the comparison imaging result indicators are shown in Table 1.

[0117] Table 1

[0118] Method TBR ENT Methods in other literature 36.39 4.912 The method according to the present invention 47.49 4.069

[0119] According to Table 1, the TBR value of the present application is 47.49, which is significantly higher than 36.39 in the method in the other literature, indicating that the present application has a significant advantage in the utilization efficiency of the time-frequency bandwidth. The improvement of TBR means that under the same bandwidth, the present application can better meet the imaging demand of wide swath.

[0120] In addition, the ENT is another indicator of imaging quality, which is used to measure the richness of image information. The lower the ENT value, the clearer the image and the less redundant information. The ENT value of the present application is 4.069, which is significantly lower than 4.912 of the method in other literatures, indicating that the method of the present application also has an advantage in imaging clarity.

[0121] The method of the present application realizes higher TBR and lower ENT by jointly optimizing the pulse emission mode and the SAR imaging algorithm. Compared with the method in other literatures, the present application significantly improves the imaging resolution and quality while realizing a wide swath.

[0122] In order to realize the above-mentioned embodiments, the present application further provides an electronic device, comprising: a processor, and a memory connected with the processor in communication; the memory stores computer execution instructions; the processor executes the computer execution instructions stored in the memory to realize the method provided by the foregoing embodiments.

[0123] In order to realize the above-mentioned embodiments, the present application further provides a computer readable storage medium, wherein the computer readable storage medium stores computer execution instructions, and the computer execution instructions are executed by a processor to realize the method provided by the foregoing embodiments.

[0124] In order to realize the above-mentioned embodiments, the present application further provides a computer program product, comprising a computer program, which is executed by a processor to realize the method provided by the foregoing embodiments.

[0125] The collection, storage, use, processing, transmission, provision and disclosure of personal information of users involved in the present application comply with the relevant laws and regulations, and do not violate public order and good customs.

[0126] It should be noted that the personal information from the user should be collected for legal and reasonable purposes, and should not be shared or sold outside these legal uses. In addition, such collection / sharing should be carried out after the user's informed consent is obtained, including but not limited to informing the user to read the user agreement / user notice before the user uses the function, and signing the agreement / authorization including authorization of relevant user information. In addition, any necessary steps should be taken to protect and ensure access to such personal information data, and to ensure that other people with access to personal information data comply with their privacy policy and processes.

[0127] The present application contemplates an implementation that provides users with the ability to selectively opt in or opt out of permitting the collection and / or use of their personal information data. That is, the present disclosure contemplates providing users with the ability to prevent or limit the collection and / or use of their personal information data. For example, the present disclosure contemplates providing users with the ability to prevent or limit the collection and / or use of their personal information data by, for example, blocking or deleting cookies. In addition, the present disclosure contemplates providing users with the ability to determine whether and how to interact with the present disclosure by, for example, blocking web beacons. Further, the present disclosure contemplates providing users with the ability to access and / or edit their personal information data when such data is collected by the present disclosure. Additionally, the present disclosure contemplates that the collection and / or use of personal information data can be done in a manner that complies with all other applicable laws and regulations, including, for example, the European Union’s General Data Protection Regulation (GDPR), the California Consumer Privacy Act (CCPA), and / or any comparable privacy laws.

[0128] In the foregoing detailed description, the description used with respect to the terms "one embodiment", "some embodiments”, "an example”, "a specific example” or "some examples” etc. means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present application. The illustrative appearances of the above-mentioned terms in the description are not necessarily referred to the same embodiment or example. Moreover, the particular features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples. Furthermore, the description herein of various embodiments or examples of the present application has been made with reference to the accompanying drawings. However, the description is not intended to limit the application to the particular examples or embodiments described. As such, the scope of the present application is to be interpreted only in conjunction with the appended claims.

[0129] In addition, the terms "first", "second", etc. are used herein only to describe various features, and do not imply relative importance or a number of the features. Thus, a feature defined with "first", "second", etc. can include at least one of the feature. In the description of the present application, the meaning of "a plurality" is at least two, for example, two, three, etc., unless otherwise specifically defined.

[0130] Any processes or methods described in the flow charts or otherwise described herein can be understood as representing code modules, segments, or portions of code that include one or more executable instructions for implementing specific logic functions (or steps) of the processes. It will be understood that the scope of the preferred embodiments of the present application encompasses also other implementations that can not be exactly as described in the flow charts, but which can include fewer, additional, or different steps, performed in a different order, including substantially concurrently or in reverse order, and which can include additional, fewer, or different components.

[0131] The logic and / or steps represented in the flowcharts and / or described herein, for example, can be considered as a sequence of executable instructions stored in a computer readable medium, which can be executed by an instruction execution system, apparatus or device, such as a computer-based system, a processor-based system, or other system that can fetch the instructions from the instruction execution system, apparatus or device and execute the instructions, or a combination thereof. For the purposes of this specification, a "computer readable medium" can be any apparatus that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus or device. The computer readable medium can specifically be, but is not limited to, the following: an electronic connection (electronic apparatus) having one or more wires, a portable computer diskette (magnetic apparatus), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, and a portable compact disk read-only memory (CDROM). In addition, the computer readable medium can even be paper or other suitable medium upon which the program can be printed, because the program can be electronically obtained, for example, by optically scanning the paper or other medium, then

[0132] It should be understood that portions of the application can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in a memory and executed by a suitable instruction execution system. As such, if implemented in hardware, and in another embodiment, any of the following technologies, known in the art, or a combination thereof, can be used: discrete logic circuitry having logic gates for implementing logic functions on data signals, application specific integrated circuits having appropriate combinational logic gates, programmable gate arrays (PGA), field programmable gate arrays (FPGA), and the like.

[0133] Those of ordinary skill in the art can understand that all or part of the steps carried out by the above-mentioned embodiment methods can be completed by programs instructing relevant hardware, and the programs can be stored in a computer readable storage medium. When the programs are executed, they include one of the steps of the method embodiments or a combination thereof.

[0134] In addition, each of the function units in each embodiment of the present application can be integrated in one processing module, or each unit can be physically present separately, or two or more units can be integrated in one module. The integrated module can be realized in the form of hardware or in the form of a software function module. When the integrated module is realized in the form of a software function module and sold or used as an independent product, it can also be stored in a computer readable storage medium.

[0135] The storage medium mentioned above can be a read-only memory, a magnetic disk or an optical disk, etc. Although the embodiments of the present application have been shown and described above, it should be understood that the above embodiments are exemplary and should not be construed as limiting the present application, and those skilled in the art can make changes, modifications, replacements and variations to the above embodiments within the scope of the present application.

[0136] It should be understood that the various forms of flow shown above can be reordered, added or deleted steps. For example, each step described in the present application can be executed in parallel, sequentially or in different order, as long as the desired results of the technical solutions of the present application can be achieved, which is not limited herein.

[0137] The above detailed description does not constitute a limitation on the scope of protection of the present application. Those skilled in the art should understand that various modifications, combinations, sub-combinations and replacements can be made according to design requirements and other factors. Any modifications, equivalent replacements and improvements within the spirit and principles of the present application should be included in the scope of protection of the present application.

Claims

1. A joint optimization SAR imaging method incorporating a learnable transmit pattern and imaging algorithm, characterized by, Includes the following steps: S1. Generate a non-uniform pseudo-echo signal based on SAR system parameters and simulation scenario, as the input signal for the imaging iteration process, and initialize the imaging result matrix as a zero matrix, as the initial state for the imaging iteration process. S2. Based on the current imaging results, first generate an approximate echo using the inverse imaging operator, calculate the residual between the approximate echo and the simulated echo, then use the imaging operator to obtain the imaging result of the residual, and use the imaging result of the residual to linearly update the imaging result of the previous round. S3. Map the updated linear imaging results to the sparse domain and use a soft thresholding iterative function to filter out noise to obtain nonlinear imaging results; S4. Repeat the linear and nonlinear update steps until the set number of iterations is reached to obtain the final imaging result; S5. After the iteration is completed, calculate the loss function and gradient, and use a random parameter freezing strategy for the pulse emission time. Randomly select a portion of the pulse emission time to freeze the gradient of the parameters. S6. Backpropagate and update the parameters of the pulse emission mode and imaging algorithm based on the gradient updated by the random parameter freezing strategy; S7. Repeat steps S2-S6 until convergence is achieved, resulting in the final optimized pulse emission mode and the matching imaging algorithm.

2. The method of claim 1, wherein, The process of generating a non-uniform pseudo-echo signal based on SAR system parameters and a simulated scenario as the input signal for the imaging iteration process, and initializing the imaging result matrix as a zero matrix as the initial state for the imaging iteration process, includes: Based on the SAR system parameters Θ and the simulation scenario X, a simulated azimuth non-uniform echo signal Y is generated as the input signal for the imaging iteration process. Initialize the imaging result X 0 is a zero matrix as the initial state of the imaging iteration process.

3. The method of claim 2, wherein, Based on the current imaging results, an approximate echo is first generated using an inverse imaging operator. The residual between the approximate echo and the simulated echo is calculated. Then, the imaging operator is used to obtain the imaging result of the residual. Finally, the imaging result of the residual is used to linearly update the previous round of imaging results, including: Based on the imaging results X of the previous round k-1 , the approximate echo matrix G(X k-1 ) is obtained according to the approximate observation inverse imaging operator, wherein G(·) represents the inverse imaging operator; computing a residual of the approximated echo matrix G(X k-1 ) and a simulated echo Y, and obtaining an approximated imaging result H(Y-G(X k-1 )) of the residual according to an imaging operator, where H(·) represents the imaging operator. The residual coefficients are multiplied with the approximated imaging result H(Y - G(X k-1 )) and added to the imaging result X k-1 of the previous round to obtain the updated linear imaging result, which is expressed as: Z k = X k-1 + a k · H(Y - G(X k-1 )) wherein Z k is the linear imaging result updated for the kth iteration.

4. The method of claim 3, wherein, The step of mapping the updated linear imaging results to a sparse domain and filtering out noise using a soft-thresholding iterative function to obtain nonlinear imaging results includes: Separate the linear imaging result Z as a complex matrix k The real and imaginary parts are used to transform it into a real matrix D(Z). k ); The real matrix D(Z k ) is mapped to a sparse domain by three-layer convolution Conv, and a soft threshold iteration function is used to filter out noise to obtain a result after filtering out noise Wherein, F k represents the processing process of three-layer convolution Conv, represents the soft threshold iteration function; The results filtered out of noise are reflected back to the image domain through three layers of convolution Conv mapping, and are used as a residual error with the initial complex matrix Z k Want to add, get the updated nonlinear imaging results, expression: wherein, represents the process of three-layer convolution Conv reflecting the result back to the image domain, C is a constant coefficient of residual adjustment, and X k is the nonlinear imaging result updated for the kth iteration.

5. The method of claim 4, wherein, The loss function for calculation includes: Computing a less-than-zero penalty term for a pulse emission pattern Penalty term for mapping of computed imaging algorithm in sparse domain to image domain computing a mean squared error, MSE, distance between the final imaged function and the target image The loss function is calculated as follows: wherein L(Γ Total ) is a loss value of the loss function, γ1, γ2, γ3 are weight coefficients of different loss terms in the loss function, is the total number of pulses in the pulse emission mode, is the emission parameter of the n a th pulse, J is the total number of imaging images, j is the number of imaging images, K is the set number of iterations, GT is the target image, X K is the final imaging result, denotes the L2 norm.

6. The method of claim 5, wherein, The random parameter freezing strategy for pulse emission moments, which randomly selects a portion of the pulse emission moments for gradient freezing of parameters, includes: A random parameter freezing strategy is applied to randomly generate a mask sequence of the same length as the pulse emission time, expressed as: Each element in the mask sequence is a random variable that follows a Bernoulli distribution; The mask sequence is multiplied with the gradient of the loss function with respect to the pulse emission pattern d η to achieve the purpose of random zeroing, expressed as: wherein, represents a loss function for the pulse emission pattern d η performs gradient calculation to obtain a corresponding gradient vector; represents multiplying the mask sequence m with the gradient vector item by item.​ 7. An electronic device, comprising: include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory to implement the method as described in any one of claims 1-6.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method as described in any one of claims 1-6.

9. A computer program product, characterised in that, Includes a computer program that, when executed by a processor, implements the method of any one of claims 1-6.

Images