Adaptive parameter adjustment method based on prior information and one-dimensional range profile peak detection algorithm

Abstract

The invention discloses a self-adaptive parameter adjustment method based on prior information and a one-dimensional range profile peak detection algorithm, and the method comprises the steps: obtaining multi-band ISAR echo signals, and carrying out the splicing of the ISAR echo signals to form sparse aperture echo signals; constructing an optimization objective function and initializing related parameters, wherein the related parameters comprise regularization parameters; the regularization parameter is used as the input of an ADMM algorithm, the sparse aperture echo signal is reconstructed, and a multi-band fusion ISAR echo is obtained; peak value detection is carried out on the one-dimensional range profile of the multi-band fusion ISAR echo, and the number of detected peak values is evaluated according to the number of known targets in prior information; and according to an evaluation result, self-adaptively adjusting related parameters until the reconstructed multi-band fused ISAR echo meets a preset requirement, and obtaining an optimal regularization parameter. The method can be well compatible with various parameters, needing to be manually adjusted, of the ADMM algorithm, and the automation degree of parameter adjustment is improved in various scenes.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of radar, and particularly relates to an adaptive parameter adjusting method based on prior information and a one-dimensional range profile peak detection algorithm. BACKGROUND

[0002] Inverse Synthetic Aperture Radar (ISAR) is a method of using the relative motion between a target and a radar to synthesize a large aperture through signal processing technology to realize high-resolution imaging. Compared with traditional radars, ISAR can obtain higher lateral resolution without physically increasing the size of the antenna. The advantages of ISAR technology mainly include strong penetration ability, robustness not limited by light and weather conditions, and sensitivity to target micro-motion characteristics, which enable it to maintain good imaging performance in complex electromagnetic environments and adverse weather. In addition, ISAR imaging does not depend on the cooperation of the target, and is suitable for long-range reconnaissance and identification of unknown or hostile targets.

[0003] Traditional ISAR imaging is based on the "rotating platform model", which regards the target as being composed of a series of independent, isotropic strong scattering points. In the imaging coordinate system (range-Doppler domain), these strong scattering points only occupy a very small part of the entire image domain, and most of the energy is zero or close to zero. This natural sparsity provides a good basis for signal sparse reconstruction.

[0004] Further, through ADMM (Alternating Direction Method of Multipliers) type algorithms, multiple sub-signals with small bandwidth can be fused and sparsely reconstructed to obtain signals with larger bandwidth, achieving resolution improvement without increasing bandwidth. Two-dimensional compressed sensing (CS) is an effective method to solve the problem of two-dimensional sparse ISAR imaging, which can decompose the joint optimization problem into sub-problems to be solved alternately, thereby realizing the fusion and reconstruction of ISAR echoes. In addition, on the basis of the conventional L1 norm method, the algorithm performance can be optimized through "weighting"; for example, there is a constrained L1 regularization alternating direction multiplier method in the prior art, which uses a convolution operator to perform convolution calculation on any pixel on the image, i.e. weighting the pixel and its surrounding 8 pixels, to realize the dilation processing of the image pixel.

[0005] However, the existing ADMM type algorithm is extremely dependent on the setting of the initial parameters when performing ISAR echo signal fusion reconstruction, needs to manually adjust the parameters according to the result of each reconstruction until the expected result appears, and thus is low in efficiency in actual use, and due to the unknown priori of the data itself, the artificial cost will also increase when the number of echoes to be processed increases. SUMMARY

[0006] In order to solve the above problems existing in the prior art, the present application provides an adaptive parameter adjustment method based on prior information and one-dimensional range profile peak detection algorithm. The technical problem to be solved by the present application is realized by the following technical scheme: In a first aspect, the present application provides an adaptive parameter adjustment method based on prior information and one-dimensional range profile peak detection algorithm, comprising: obtaining ISAR echo signals of multiple frequency bands, and splicing to form a sparse aperture echo signal; constructing an optimization objective function and initializing related parameters, wherein the related parameters include a regularization parameter; taking the regularization parameter as an input of an alternating direction multiplier ADMM type algorithm, reconstructing the sparse aperture echo signal, and obtaining a multi-band fused ISAR echo; performing peak detection on a one-dimensional range profile of the multi-band fused ISAR echo, and evaluating the number of detected peaks according to the number of known targets in the prior information; according to the evaluation result, adaptively adjusting the related parameters until the reconstructed multi-band fused ISAR echo meets a preset requirement, and obtaining an optimal regularization parameter.

[0007] In an embodiment of the present application, the sparse aperture echo signal is represented as: ; In the formula, represents a radar cross section, represents the speed of light, represents a carrier frequency, represents an instantaneous slant range, represents an ordered subset of is a sub-pulse number, is an ordered subset of is an echo number, represents the number of stepped frequencies emitted at the observation time, represents the number of effective observations on the target within the coherent accumulation time.

[0008] In an embodiment of the present application, the step of constructing an optimization objective function and initializing related parameters comprises:​ the sparse aperture echo signal After discretization, it is expressed in matrix form as: ; In the formula, denotes the distance dictionary, denotes the multi-band fusion ISAR echo to be reconstructed, denotes the azimuth dictionary, denotes the complex noise matrix, denotes the transpose; An optimization objective function is constructed: ; In the formula, , are respectively , vector form, denotes the regularization parameter, , denotes the Kronecker product; An ADMM type algorithm is used to iteratively solve the optimization objective function, by introducing an auxiliary variable , and the augmented Lagrangian function of the optimization objective function is decomposed into three sub-problems for optimization; Among them, the augmented Lagrangian function of the optimization objective function is expressed as: ; The three sub-problems are respectively expressed as: ; ; ; In the formula, denotes the current iteration number, , denotes the Lagrange multiplier, denotes the penalty parameter, denotes the update step of the Lagrange multiplier, denotes the conjugate transpose, denotes the soft threshold operator; The regularization parameter , the left boundary , the right boundary and the evaluation index are initialized.

[0009] In an embodiment of the present application, the step of performing peak detection on the one-dimensional range image of the multi-band fusion ISAR echo and evaluating the number of detected peaks according to the number of targets known in the prior information comprises: Peak value detection and sorting are performed on the one-dimensional range image of the multi-band fused ISAR echo to obtain a peak value vector , respectively represent the first peak values in descending order; The number of targets known in prior information is compared with the number of detected peak values . .

[0010] In an embodiment of the present application, when , the step of adaptively adjusting the related parameters according to the evaluation result comprises: when , it indicates that the regularization parameter is too large, and the target information is filtered out, the regularization parameter and the right boundary are adjusted by using a first strategy. The updated regularization parameter is further used as an input of an ADMM algorithm, and the step of reconstructing the sparse aperture echo signal to obtain the multi-band fused ISAR echo is returned.

[0011] In an embodiment of the present application, the step of adjusting the regularization parameter and the right boundary by using the first strategy comprises: The regularization parameter and the right boundary are updated according to the following formula: . . In the formula, the updated right boundary is represented by .

[0012] In an embodiment of the present application, when , the step of adaptively adjusting the related parameters according to the evaluation result comprises: when , the evaluation index of the multi-band fused ISAR echo is calculated under the regularization parameter , which represents the ratio of the minimum peak value to the maximum peak value in the peak values formed by the target; if the evaluation index is less than a first preset threshold, it is determined that the regularization parameter is improperly set, which causes the sidelobe to be misjudged as the main lobe, the regularization parameter and the right boundary are adjusted by using the first strategy, and the updated regularization parameter is used as an input of an ADMM algorithm.As an input of the ADMM algorithm, return the step of reconstructing the sparse aperture echo signal to obtain the multi-band fused ISAR echo. If the evaluation index is not less than a first preset threshold, then the regularization parameter is adjusted using a second strategy, and the evaluation index of the multi-band fused ISAR echo is further calculated. , indicating the ratio of the highest side lobe to the minimum peak value formed by the target. If the evaluation index is greater than a second preset threshold, it is determined that the side lobe is too high, and the regularization parameter and the left boundary are adjusted using the second strategy. The updated regularization parameter is used as an input of the ADMM algorithm, and the step of reconstructing the sparse aperture echo signal to obtain the multi-band fused ISAR echo is returned. Conversely, it indicates that the regularization parameter is the optimal regularization parameter. In an embodiment of the present application, the step of adjusting the regularization parameter and the left boundary using the second strategy comprises: . In the formula, indicates the updated left boundary.

[0013] Compared with the prior art, the present application has the following advantages: 1. The adaptive parameter adjustment method provided by the present application is suitable for various parameters involved in the ADMM algorithm that need to be manually adjusted, has good universality, can avoid repeated manual adjustment in various scenarios, and improves the efficiency of the algorithm.

[0014] 2. The present application combines the known number of targets in the prior information, the peak information in the one-dimensional range profile of the multi-band fused ISAR echo, and the mathematical information of the soft threshold operator to design the parameter adaptive logic, which can well accommodate various parameters involved in the ADMM algorithm that need to be manually adjusted, and improve the automation degree of parameter adjustment in various scenarios.

[0015] 3. The evaluation index introduced in the present application well combines prior information and signal characteristics, and can assist in judging the quality of the output result of the ADMM algorithm and feedback adjustment.

[0016] The present application will be further described in detail below with reference to the accompanying drawings and embodiments. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1is a flow chart of the adaptive parameter adjustment method based on prior information and one-dimensional range profile peak detection algorithm provided by the embodiment of the present application; Figure 2 is a schematic diagram of the adaptive parameter adjustment method based on prior information and one-dimensional range profile peak detection algorithm provided by the embodiment of the present application. Figure 3 is a simulation result schematic diagram provided by the embodiment of the present application. DETAILED DESCRIPTION

[0018] The present application will be further described in detail below in combination with specific embodiments, but the embodiments of the present application are not limited thereto.

[0019] Figure 1 is a flow chart of the adaptive parameter adjustment method based on prior information and one-dimensional range profile peak detection algorithm provided by the embodiment of the present application, Figure 2 is a schematic diagram of the adaptive parameter adjustment method based on prior information and one-dimensional range profile peak detection algorithm provided by the embodiment of the present application. As Figures 1-2 indicated, the embodiment of the present application provides an adaptive parameter adjustment method based on prior information and one-dimensional range profile peak detection algorithm, comprising: S1, acquiring ISAR echo signals of multiple frequency bands, and splicing to form a sparse aperture echo signal.

[0020] Specifically, the ISAR signals observed under different frequency bands are spliced to form a sparse aperture echo signal: ; In the formula, denotes a radar cross section, denotes a light speed, denotes a carrier frequency, denotes an instantaneous slant range, denotes an ordered subset of is a sub-pulse number, is an ordered subset of is an echo number, denotes a number of stepped frequency waveforms transmitted when observing, denotes a number of effective observations completed on the target within a coherent accumulation time.

[0021] The sparse aperture echo signal can be regarded as a waveform of stepped frequency transmitted when observing the target continuously, and only effective observations are completed on the target within a coherent accumulation time to form an echo, wherein .

[0022] S2, construct an optimization objective function and initialize related parameters, including regularization parameters.

[0023] First, the sparse aperture echo signal is discretized and expressed in matrix form: ; wherein, , represents a distance dictionary, represents the number of sub-pulses under sparse reconstruction, represents the number of sub-pulses under ideal conditions, i.e., represents sparse band observation, represents the multi-band fusion ISAR echo to be reconstructed, represents an azimuth dictionary, , represents a complex noise matrix, represents transposition.

[0024] For the convenience of subsequent formula derivation, the matrix form is further converted into a vector form: ; wherein, , , , , represents a Kronecker product, represents stacking a matrix into a vector form by column.

[0025] In order to solve , the optimization objective function is constructed in the form of the following formula or its variants: ; wherein, represents a regularization parameter.

[0026] The above formula is solved by using an ADMM algorithm, and an auxiliary variable is introduced, and the optimization objective function is further expressed as: ; Correspondingly, its augmented Lagrangian function is: ; The augmented Lagrangian function can be decomposed into the following three sub-problems for solving: ; ; ; wherein, Represents the Lagrange multipliers. Indicates the penalty parameter. , This represents the update step size of the Lagrange multipliers, which defaults to 1. The soft threshold operator is defined as follows:

[0027] ; This represents the variable that requires a soft thresholding operation during the iteration process. This indicates a soft threshold.

[0028] For the above optimization objective function, the parameters , and All of these are determined by the properties of the echo itself, but , The parameters are merely empirical values ​​set manually. These parameters are not updated during the iteration process of ADMM-type algorithms, yet they still affect the iteration process. This makes the final result of ADMM-type algorithms highly dependent on these parameters. From the formula... It can be seen from this that if Setting it too high will cause the reconstructed signal to be excessively suppressed, potentially resulting in the loss of important information; conversely, if... Setting the value too low will result in an excessive amount of clutter.

[0029] Further, initialize the regularization parameters. Left boundary Right boundary And evaluation indicators.

[0030] S3. The regularization parameter is used as the input of the alternating direction multiplier ADMM-type algorithm to reconstruct the sparse aperture echo signal and obtain multi-band fused ISAR echo.

[0031] S4. Perform peak detection on the one-dimensional range profile of the multi-band fused ISAR echo, and evaluate the number of detected peaks based on the known number of targets in the prior information.

[0032] It should be noted that ADMM-type algorithms are used to iteratively solve the objective function by introducing auxiliary variables. Before decomposing the objective function into three sub-problems for optimization, the regularization parameters are initialized. Left boundary Right boundary And evaluation indicators: , .

[0033] Specifically, in step S4, peak detection and sorting are performed on the one-dimensional range image of the multi-band fused ISAR echo to obtain a peak vector , respectively represent the first peak values in descending order; then, the number of targets known in prior information is compared with the number of detected peaks .

[0034] S5, according to the evaluation result, adaptively adjusting the related parameters until the reconstructed multi-band fused ISAR echo meets the preset requirements, and obtaining the optimal regularization parameter.

[0035] When , according to the evaluation result, the step of adaptively adjusting the related parameters comprises: When , it indicates that the regularization parameter is too large, and the target information is filtered out, and the regularization parameter and the right boundary are adjusted by using the first strategy. Further, the updated regularization parameter is taken as the input of the ADMM algorithm, and the step of reconstructing the sparse aperture echo signal to obtain the multi-band fused ISAR echo is returned.

[0036] Exemplarily, the step of adjusting the regularization parameter and the right boundary by using the first strategy comprises: The regularization parameter and the right boundary are updated according to the following formula: ; In the formula, the updated right boundary is represented by .

[0037] On the other hand, when , according to the evaluation result, the step of adaptively adjusting the related parameters comprises: When , under the regularization parameter , the evaluation index of the multi-band fused ISAR echo is calculated, which represents the ratio of the minimum peak value to the maximum peak value in the peaks formed by the target; if the evaluation index is less than a first preset threshold, it is determined that the regularization parameter is improperly set, which causes the sidelobe to be misjudged as the main lobe, and the regularization parameter and the right boundary are adjusted by using the first strategy.the updated regularization parameter as the input of the ADMM algorithm, reconstructing the sparse aperture echo signal to obtain the multi-band fused ISAR echo. if the evaluation index is not less than a first preset threshold, then the regularization parameter is further calculated, and the evaluation index of the multi-band fused ISAR echo is calculated, wherein the evaluation index represents the ratio of the highest side lobe to the minimum peak value formed by the target. if the evaluation index is greater than a second preset threshold, it is determined that the side lobe is too high, and the regularization parameter and the left boundary are adjusted by using the second strategy, and the updated regularization parameter is used as the input of the ADMM algorithm, and the step of reconstructing the sparse aperture echo signal to obtain the multi-band fused ISAR echo is returned. if the evaluation index

[0038] is less than the first preset threshold, it is determined that the regularization parameter is not optimal. the step of adjusting the regularization parameter and the left boundary includes: ; in the formula, the updated left boundary is represented.

[0039] The adaptive parameter adjustment method based on the prior information and the one-dimensional range profile peak detection algorithm provided by the present application is further described below through simulation experiments.

[0040] Figure 3 is a simulation result schematic diagram provided by an embodiment of the present application. As shown in Figure 3 , the regularization parameter is determined as 0.0625 after 4 rounds of parameter adjustment, which means that under the same conditions, manual parameter adjustment needs to run the program repeatedly 4 times to obtain the same result, and therefore the adaptive parameter adjustment algorithm undoubtedly saves the labor cost.

[0041] From the above embodiments, it can be seen that the present application has the following beneficial effects: 1. The adaptive parameter adjustment method provided by the present application is suitable for various parameters involved in the ADMM algorithm and needs to be manually adjusted, has good universality, can avoid repeated manual adjustment of the parameters in various scenarios, and improves the efficiency of the algorithm.

[0042] 2、The application combines the known target number in prior information, peak information in one-dimensional range image of multi-band fusion ISAR echo and mathematical information of soft threshold operator to design parameter adaptive logic, which can well compatible with various parameters of ADMM algorithm that need to be manually adjusted, and improve the automation degree of parameter adjustment in various scenarios.

[0043] 3、The evaluation index introduced in the application well combines prior information and signal characteristics, which can assist in judging the quality of ADMM algorithm output results and feedback adjustment.

[0044] In the description of the application, the description of the terms "one embodiment", "some embodiments", "example", "specific example" or "some examples" means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the application. In the specification, the illustrative description of the above terms is not necessarily for the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner. In addition, those skilled in the art can combine and combine different embodiments or examples described in the specification.

[0045] Although the application is described herein in connection with various embodiments, other variations to the disclosed embodiments can be understood and effected by those skilled in the art from a study of the drawings, the disclosure, and the appended claims in implementing the claimed application.

[0046] The above is a further detailed description of the application in combination with specific preferred embodiments, and the specific implementation of the application cannot be limited to these descriptions. For those skilled in the art to which the application belongs, without departing from the concept of the application, a number of simple deductions or substitutions can be made, which should be regarded as falling within the protection scope of the application.

Claims

1. An adaptive parameter tuning method based on prior information and a one-dimensional distance image peak detection algorithm, characterized in that, include: Multi-band ISAR echo signals are acquired and spliced ​​to form sparse aperture echo signals; Construct an optimization objective function and initialize relevant parameters, including regularization parameters; The regularization parameter is used as the input of an alternating direction multiplier (ADMM) algorithm to reconstruct the sparse aperture echo signal and obtain multi-band fused ISAR echo. Peak detection is performed on the one-dimensional range profile of the multi-band fused ISAR echo, and the number of detected peaks is evaluated based on the number of targets known in the prior information. Based on the evaluation results, the relevant parameters are adaptively adjusted until the reconstructed multi-band fused ISAR echo meets the preset requirements, thus obtaining the optimal regularization parameters.

2. The adaptive parameter tuning method based on prior information and a one-dimensional distance image peak detection algorithm according to claim 1, characterized in that, The sparse aperture echo signal is represented as follows: ； In the formula, Indicates radar cross-section, Represents the speed of light. Indicates the carrier frequency. Indicates instantaneous slant distance. express An ordered subset, The sub-pulse number. for An ordered subset, Echo sequence number This indicates the number of step frequencies emitted during observation. This indicates the number of effective observations completed for the target within the coherent accumulation time.

3. The adaptive parameter tuning method based on prior information and a one-dimensional distance image peak detection algorithm according to claim 1, characterized in that, The steps for constructing the objective function and initializing the relevant parameters include: The sparse aperture echo signal Discretized, it can be represented in matrix form: ； In the formula, Distance dictionary, This represents the ISAR echo from the multi-band fusion process to be reconstructed. A dictionary representing directions. Represents the complex noise matrix. Indicates transpose; Construct the optimization objective function: ； In the formula, , They are respectively , vector form, Represents the regularization parameter. , Indicates the Kronecker product; The objective function is solved iteratively using an ADMM-type algorithm, by introducing auxiliary variables. The augmented Lagrange function of the objective function is decomposed into three sub-problems for optimization. The augmented Lagrangian function of the objective function is expressed as: ； The three sub-problems are respectively represented as follows: ； ； ； In the formula, Indicates the current iteration number. , Represents the Lagrange multipliers. Indicates the penalty parameter. This represents the update step size of the Lagrange multipliers. This indicates the conjugate transpose. This represents the soft threshold operator; Initialize regularization parameters Left boundary Right boundary And evaluation indicators.

4. The adaptive parameter tuning method based on prior information and a one-dimensional distance image peak detection algorithm according to claim 3, characterized in that, The steps of performing peak detection on the one-dimensional range profile of the multi-band fused ISAR echoes and evaluating the number of detected peaks based on the known number of targets in the prior information include: Peak detection and sorting are performed on the one-dimensional range profile of the multi-band fused ISAR echoes to obtain the peak vector. , These represent the order of the descending order of the first... One peak; Compare the number of known targets in the prior information. With the number of detected peaks .

5. The adaptive parameter tuning method based on prior information and a one-dimensional distance image peak detection algorithm according to claim 4, characterized in that, when The step of adaptively adjusting the relevant parameters based on the evaluation results includes: when When, it indicates the regularization parameter. If the value is too large and target information is filtered out, adjust the regularization parameters using the first strategy. and the right boundary Further update the regularization parameters As input to ADMM-type algorithms, the steps of reconstructing the sparse aperture echo signal to obtain multi-band fused ISAR echoes are returned.

6. The adaptive parameter tuning method based on prior information and a one-dimensional distance image peak detection algorithm according to claim 5, characterized in that, Adjusting regularization parameters using the first strategy and the right boundary The steps include: Update the regularization parameters according to the following formula. and right boundary : ； ； In the formula, This indicates the updated right boundary.

7. The adaptive parameter tuning method based on prior information and a one-dimensional distance image peak detection algorithm according to claim 4, characterized in that, when The step of adaptively adjusting the relevant parameters based on the evaluation results includes: when At that time, in the regularization parameter The following steps are used to calculate the evaluation index of the multi-band fused ISAR echo. , representing the ratio of the minimum peak value to the maximum peak value among the peak values ​​formed by the target; if the evaluation index If the value is less than the first preset threshold, then the regularization parameter is determined. Improper settings caused the sidelobe to be misidentified as the main lobe. The first strategy was used to adjust the regularization parameters. and right boundary The updated regularization parameters As input to ADMM-type algorithms, the steps of reconstructing the sparse aperture echo signal to obtain multi-band fused ISAR echoes are returned. If evaluation indicators If it is not less than the first preset threshold, then the regularization parameter... Next, the evaluation index of the multi-band fused ISAR echo is further calculated. , which represents the ratio of the highest sidelobe to the minimum peak value formed by the target; If evaluation indicators If the sidelobe is greater than the second preset threshold, it is determined that the sidelobe is too high, and the regularization parameter is adjusted using the second strategy. and left boundary The updated regularization parameters As input to ADMM-type algorithms, the returned step is to reconstruct the sparse aperture echo signal to obtain a multi-band fused ISAR echo; otherwise, it indicates that the regularization parameter... This is the optimal regularization parameter.

8. The adaptive parameter tuning method based on prior information and a one-dimensional distance image peak detection algorithm according to claim 7, characterized in that, Adjusting regularization parameters using the second strategy and left boundary The steps include: ； ； In the formula, This indicates the updated left boundary.

Images