Target identification method and device based on azimuth angle geometric constraint classifier HRRP

Abstract

The invention relates to a target identification method and device based on an azimuth angle geometric constraint classifier HRRP. The method comprises the following steps: constructing a radar target recognition model; and inputting the unbalanced HRRP data set into a feature extraction module to extract target features, and inputting the target features into an azimuth geometric constraint classifier to perform category learning. Training the dominant category features to obtain dominant category azimuth angle frame classification sub-prototypes, and calculating sub-prototype displacement between the azimuth angle frame classification sub-prototypes of the current dominant category features and the next dominant category features by adopting a layer of convolution; and training the edge category features according to the sub-prototype displacement to obtain an edge category azimuth angle frame classification sub-prototype. And reconstructing a loss function according to the dominant category azimuth angle frame classification sub-prototype and the edge category azimuth angle frame classification sub-prototype, and then optimizing a radar target recognition model. By adopting the method, the radar target recognition performance can be improved under the condition of class imbalance.

Description

Technical Field

[0001] The present invention relates to the field of radar target recognition technology, and in particular to a HRRP target recognition method and device based on an azimuth geometric constraint classifier. Background Art

[0002] Radar can work stably under non-ideal detection conditions and has the advantage of being non-contact. Therefore, radar has become an important sensor in the field of automatic target recognition. High-resolution range profile (HRRP) can reflect the distribution of scattering points of the target along the radar line of sight, and contains many physical characteristics of the target, such as the distribution of scattering points, the position and intensity of the scattering center, and the size of the target. Therefore, radar target recognition methods based on HRRP have attracted more and more attention from researchers in recent years and have become a hot issue in the field of radar target recognition.

[0003] At present, HRRP target recognition methods can be roughly divided into two categories: one is the traditional HRRP target recognition method, and the other is the HRRP target recognition method based on deep learning. The traditional HRRP target recognition method establishes a clear signal model or physical model, extracts the separable features of the HRRP target and designs the corresponding classifier to achieve effective recognition of the target. The deep learning-based method constructs an artificial neural network, designs a clever structure, constructs an end-to-end deep neural network, and optimizes the parameters of the neural network by using a large amount of labeled training data to train and optimize the parameters of the neural network, thereby obtaining a set of neural network parameters under the background of big data. The trained parameters are combined with the constructed neural network to achieve effective recognition of the target. Although these methods have achieved good recognition performance, most of them are trained on a class-balanced data set, that is, each class of target has the same number of training samples. However, in practical applications, especially for non-cooperative targets, due to the changes in observation conditions, the target's motion trajectory, and the time when the target appears, it is difficult or even impossible to obtain the same number of training samples for each class of target. The data set that can be obtained for training is often class-imbalanced. In the class-imbalanced data set, the class with sufficient number of samples is called the dominant class, and the class with insufficient number of samples is called the marginal class. Directly applying the above method based on the assumption of a balanced dataset to HRRP target recognition under imbalanced conditions will cause the model to overfit to the dominant category and ignore the learning of marginal categories, resulting in a decrease in model performance and even the risk of failure. Summary of the invention

[0004] Based on this, it is necessary to provide a HRRP target recognition method and device based on azimuth geometric constraint classifier, which can improve the radar's target recognition performance under the condition of class imbalance data, in order to solve the above technical problems.

[0005] A HRRP target recognition method based on an azimuth geometric constraint classifier, the method comprising: The radar target recognition model is constructed based on HRRP. The radar target recognition model includes: feature extraction module, azimuth geometric constraint classifier and loss function.

[0006] The unbalanced HRRP dataset is input into the feature extraction module to extract the target features, and the target features are input into the azimuth geometric constraint classifier for category learning. The target features include: dominant category features and marginal category features.

[0007] The dominant category features are trained to obtain the dominant category azimuth frame classification sub-prototype, and a layer of convolution is used to calculate the sub-prototype displacement between the azimuth frame classification sub-prototype of the current dominant category feature and the next dominant category feature.

[0008] The edge category features are trained according to the sub-prototype displacement to obtain the edge category azimuth frame classification sub-prototype.

[0009] The loss function is reconstructed according to the dominant category azimuth frame classification sub-prototype and the marginal category azimuth frame classification sub-prototype to obtain the overall loss function. The radar target recognition model is optimized according to the overall loss function. The unbalanced HRRP dataset is trained with the optimized radar target recognition model to output the target image.

[0010] A HRRP target recognition device based on an azimuth geometric constraint classifier, the device comprising: The model building module is used to build a radar target recognition model based on HRRP. The radar target recognition model includes: feature extraction module, azimuth geometric constraint classifier and loss function.

[0011] The feature learning module is used to input the unbalanced HRRP data set into the feature extraction module to extract the target features, and input the target features as the training sample data set into the azimuth geometric constraint classifier for category learning. The training sample data set includes: dominant category features and marginal category features.

[0012] The sub-prototype displacement acquisition module is used to train the dominant category features to obtain the dominant category azimuth frame classification sub-prototype, and use a layer of convolution to calculate the sub-prototype displacement between the azimuth frame classification sub-prototype of the current dominant category feature and the next dominant category feature.

[0013] The edge category feature training module is used to train the edge category feature according to the sub-prototype displacement to obtain the edge category azimuth frame classification sub-prototype.

[0014] The target image acquisition module is used to reconstruct the loss function according to the dominant category azimuth frame classification sub-prototype and the marginal category azimuth frame classification sub-prototype to obtain the overall loss function, optimize the radar target recognition module according to the overall loss function, train the unbalanced HRRP data set through the optimized radar target recognition module, and output the target image.

[0015] The above-mentioned HRRP target recognition method and device based on azimuth geometric constraint classifier, first, the feature extraction module uses a deep convolutional neural network to perform multi-scale feature mining on the original HRRP data, suppresses the noise interference of the dominant category through an adaptive feature selection mechanism, and retains the azimuth sensitive features of the target. In the design of the azimuth geometric constraint classifier, an innovative dual-branch training strategy based on sub-prototype displacement is proposed: in view of the fact that the dominant category samples are sufficient, a dominant category azimuth frame classification sub-prototype library is constructed, and the sub-prototype displacement vector between adjacent frames is calculated through a dynamic convolution layer to capture the geometric change law of the target movement; for the edge category, the displacement vector learned by the dominant category is used as a priori constraint, and the edge category azimuth frame classification sub-prototype is generated through geometric transformation. This improves the generalization performance of the model for edge category recognition without reducing the recognition performance of the model for the dominant category, and solves the problem of inaccurate feature distribution under small sample conditions. By constructing a multi-task optimization system including prototype contrast loss, displacement constraint loss and category balance loss, while maintaining the dominant category discrimination ability, the model is forced to learn the geometric mapping relationship between the edge category features and the dominant category displacement pattern. This further improves the recognition performance of the model under the condition of imbalanced categories. The overall loss function of the model consists of classification loss and decoupling loss, which can enhance the recognition accuracy of different category features under the condition of imbalanced data and improve the radar's recognition performance of targets. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A flowchart of a HRRP target recognition method based on an azimuth geometric constraint classifier in one embodiment; Figure 2 Flow chart of the HRRP target recognition steps based on the azimuth geometric constraint classifier in one embodiment Figure 3 is a diagrammatic illustration of an azimuth geometry constraint classifier in one embodiment; Figure 4 HRRP images of three different types of aircraft measured in one embodiment, where (a) is an An-26 aircraft, (b) is a propeller-type aircraft, and (c) is a Yak-42 aircraft; Figure 5It is a schematic diagram of the recognition accuracy rate obtained on a measured aircraft data set in one embodiment; Figure 6 A schematic diagram of a confusion matrix of recognition accuracy under class imbalance conditions on a measured aircraft dataset in one embodiment; Figure 7 It is a structural block diagram of a HRRP target recognition device based on an azimuth geometric constraint classifier in one embodiment. DETAILED DESCRIPTION

[0017] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0018] In one embodiment, Figure 1 As shown, a HRRP target recognition method based on an azimuth geometric constraint classifier is provided, comprising the following steps: Step 102: construct a radar target recognition model based on HRRP.

[0019] The radar target recognition model includes: feature extraction module, azimuth geometry constraint classifier and loss function.

[0020] Step 104: Input the unbalanced HRRP data set into a feature extraction module to extract target features, and input the target features into an azimuth geometric constraint classifier for category learning.

[0021] Target features include: dominant category features and marginal category features.

[0022] Step 106, training the dominant category feature to obtain a dominant category azimuth frame classification sub-prototype, and using a layer of convolution to calculate the sub-prototype displacement between the azimuth frame classification sub-prototype of the current dominant category feature and the next dominant category feature.

[0023] Step 108: training edge category features according to the sub-prototype displacement to obtain edge category azimuth frame classification sub-prototype.

[0024] Step 110, reconstruct the loss function according to the dominant category azimuth frame classification sub-prototype and the marginal category azimuth frame classification sub-prototype to obtain the overall loss function, optimize the radar target recognition model according to the overall loss function, train the unbalanced HRRP data set with the optimized radar target recognition model, and output the target image.

[0025] In the above HRRP target recognition method based on azimuth geometric constraint classifier, first, the feature extraction module uses deep convolutional neural network to mine multi-scale features of the original HRRP data, and suppresses the noise interference of the dominant category through adaptive feature selection mechanism, retaining the azimuth sensitive features of the target. In the design of azimuth geometric constraint classifier, an innovative dual-branch training strategy based on sub-prototype displacement is proposed: in view of the fact that the dominant category samples are sufficient, a dominant category azimuth frame classification sub-prototype library is constructed, and the sub-prototype displacement vector between adjacent frames is calculated through the dynamic convolution layer to capture the geometric change law of target motion; for edge categories, the displacement vector learned by the dominant category is used as a priori constraint, and the edge category azimuth frame classification sub-prototype is generated through geometric transformation. This improves the generalization performance of the model for edge category recognition without reducing the recognition performance of the model for the dominant category, and solves the problem of inaccurate feature distribution under small sample conditions. By constructing a multi-task optimization system including prototype contrast loss, displacement constraint loss and category balance loss, while maintaining the dominant category discrimination ability, the model is forced to learn the geometric mapping relationship between edge category features and dominant category displacement patterns. This further improves the recognition performance of the model under the condition of imbalanced categories. The overall loss function of the model consists of classification loss and decoupling loss, which can enhance the recognition accuracy of different category features under the condition of imbalanced data and improve the radar's recognition performance of targets.

[0026] In one embodiment, the unbalanced HRRP data set is input into the feature extraction module, and the long short-term memory network is used to extract the change characteristics of the target strong scattering center in the unbalanced HRRP data set to obtain the target feature: ; in, is the target feature, The unbalanced HRRP data set is used as the input to the azimuth geometry constraint classifier for category learning. The unbalanced HRRP dataset includes: dominant category data and marginal category data.

[0027] In one embodiment, the target azimuth frame classification sub-prototype corresponding to the target feature is learned by the azimuth geometric constraint classifier, and the dominant category feature is trained to obtain the dominant category azimuth frame classification sub-prototype. The geometric constraint is defined according to the learned dominant category azimuth frame classification sub-prototype: ; in, is the set of sub-prototype displacements of the dominant category azimuth frame classification, is the number of learned target azimuth frame classification sub-prototypes, is the number of dominant categories. A layer of convolution is used to calculate the sub-prototype displacement between the azimuth frame classification sub-prototype of the current dominant category feature and the next dominant category feature: ; in, For the The leading category The dominant category azimuth frame classification sub-prototype and the The sub-prototype displacement between the azimuth frame classification sub-prototypes, For the transformation operation, For the The dominant category is learned through training dominant category azimuth frame classification sub-prototypes, For the The dominant category is learned through training Azimuth frame classification sub-prototypes.

[0028] In one embodiment, edge category features are trained using an azimuth geometry constraint classifier based on the sub-prototype displacement: ; in, For the The edge categories newly generated edge category azimuth frame classification sub-prototype, For the The edge categories are learned through the training sample dataset. edge category azimuth frame classification sub-prototypes, is the total number of learned edge category azimuth frame classification sub-prototypes, For the The leading category The dominant category azimuth frame classification sub-prototype and the The sub-prototype displacement between the azimuth frame classification sub-prototypes, For averaging operation, , .

[0029] In one embodiment, the azimuth geometry constraint classifier uses cosine similarity as a distance metric to classify the target feature into the category of the classification sub-prototype that is closest to the current target feature: ; in, is the cosine similarity, is the number of target feature categories, is the total number of categories, is the total number of dominant categories, is the total number of edge categories, is the target feature, For the Category Azimuth frame classification sub-prototypes, is the distance measure.

[0030] In one embodiment, the probability that the target feature belongs to the target azimuth frame classification sub-prototype is generated based on the dominant category azimuth frame classification sub-prototype, the marginal category azimuth frame classification sub-prototype, and the current target feature: ; in, is the probability that the target feature belongs to the target azimuth frame classification sub-prototype, is the number of learned target azimuth frame classification sub-prototypes, For the Category Azimuth frame classification sub-prototypes; the classification loss generated by the azimuth set constraint classifier is reconstructed according to the probability: ; in, is the reconstructed classification loss function, is the HRRP data to be identified, The target category label identified by the model, For the Category Azimuth frame classification sub-prototypes.

[0031] The decoupled loss function is constructed based on the distance metric and the dominant category azimuth frame classification sub-prototype: ; in, To decouple the loss function, is the distance metric, For the Category Azimuth frame classification sub-prototypes, is the distance metric The transpose of . The overall loss function is constructed based on the classification loss function and the decoupling loss function: ; in, is the overall loss function, The radar target recognition model is optimized through the overall loss function, and the unbalanced HRRP dataset is trained with the optimized radar target recognition model to output the target image.

[0032] In one of the embodiments, the azimuth geometry constraint classifier performs classification according to an intuitive principle, that is, given a HRRP sample, it is classified into the category of the azimuth prototype that is closest to it. Specifically, different classification sub-prototypes are assigned to different azimuth frames of each category. Under the condition of class imbalance, the training samples of the dominant category are more sufficient than the training samples of the marginal category. Therefore, the classification sub-prototypes of different azimuth frames of the dominant category can often be learned more effectively. However, due to the lack of training samples, the learned azimuth frame classification sub-prototypes of the marginal category are often incomplete. To address this problem, the azimuth geometry constraint classifier transfers the geometric constraints between the classification sub-prototypes of different azimuth frames of the dominant category to the classification sub-prototypes of different azimuth frames of the marginal category, so that the azimuth frame classification sub-prototypes that are difficult to learn due to the lack of training samples can be learned for the marginal category. Figure 3 A graphical illustration of the azimuth geometry constraint classifier is given.

[0033] In a class-imbalanced dataset, assuming there is The leading categories For the dominant category, we can learn For edge categories, only the target azimuth frame classification prototypes can be learned from the training samples. The geometric constraint is defined as the set of displacements of all classification sub-prototypes of the dominant category, which can be expressed as The displacement of the classified prototype can be calculated by formula (1), where Indicates The leading category The prototype and The prototype displacement between sub-prototypes, represents the transformation operation. This method uses a convolutional layer to perform the transformation operation, which enhances the displacement of the original sub-prototype. Separability and class independence in transformation space.

[0034] (1) Using the sub-prototype displacement obtained by formula (1), the newly generated azimuth frame classification sub-prototype of the edge category can be expressed as formula (2), where Indicates The azimuth frame classification sub-prototypes newly generated by edge categories, Indicates The edge categories are learned through training samples. Azimuth frame classification sub-prototypes, Represents an averaging operation.

[0035] (2) Using formula (1) and formula (2), we can obtain the classification sub-prototypes of all target azimuth frames of the dominant category and the marginal category. Based on the obtained classification sub-prototypes, the azimuth geometric constraint classifier classifies the test sample into the category of the classification sub-prototype closest to it, as shown in formula (3), where Indicates testing HRRP. Indicates Category Azimuth frame classification sub-prototypes, , represents the total number of categories, express The leading categories express Edge categories, The present invention adopts cosine similarity as the distance metric, and the cosine similarity can be expressed as formula (4).

[0036] (3) (4) in, Indicates testing HRRP Belongs to the target azimuth frame classification sub-prototype In order to satisfy the property that the sum of probabilities is 1 and the value is non-negative, the probability It can be expressed as formula (5).

[0037] (5) Based on probability , probability It can be expressed as formula (6). The classification loss function generated by the azimuth geometric constraint classifier can be expressed as formula (7), where represents the classification loss generated by the azimuth geometry constraint classifier, Indicates the sample size.

[0038] (6) (7) In one of the embodiments, in order to avoid the problem of confusion of category-related information that may occur when generating new azimuth frame classification sub-prototypes, a decoupling loss function is further proposed to ensure that the geometric constraints do not contain category-related information of the dominant category. The decoupling loss function aims to separate the geometric constraints from the category-related information of the dominant class. Therefore, in order to maximize the degree of separation between the geometric constraints and the azimuth frame classification sub-prototype of the dominant category, an orthogonal constraint is used to strengthen the decoupled learning of the geometric constraints and the azimuth frame classification sub-prototype of the dominant category. Specifically, the cosine similarity is used to measure the geometric constraints. and the dominant category azimuth frame classification sub-prototype The decoupling loss can be expressed as formula (8), where represents the decoupling loss.

[0039] (8) Furthermore, the overall loss function of the model consists of two parts: the classification loss function generated by the azimuth geometry constraint classifier and the decoupling loss function. As shown in formula (7), the decoupled loss function As shown in formula (8), the overall loss function of the model is As shown in formula (10), Represents a trade-off parameter.

[0040] (9) It is worth mentioning that an azimuth geometry constraint classifier is proposed. Under the condition of class imbalance, the azimuth geometry constraint classifier learns a classification sub-prototype for each azimuth frame of the dominant category, and uses the geometric constraints of each azimuth frame classification sub-prototype of the dominant category to generate azimuth frame classification sub-prototypes for marginal categories that have not been learned due to insufficient training samples. The azimuth geometry constraint classifier uses the distance between HRRP features and azimuth frame classification sub-prototypes in the feature space to achieve target classification and recognition. On this basis, in order to remove the category-related information of the dominant category that may be contained in the geometric constraints, a decoupling loss function is further proposed. The decoupling loss function is used to enhance the category independence of the geometric constraints. The overall loss function of the model consists of the classification loss function and the decoupling loss function generated by the azimuth geometry constraint classifier. HRRP target recognition experiments under class imbalance conditions were carried out on three types of measured aircraft data sets. The experimental results prove the effectiveness of the proposed method under class imbalance conditions and have important engineering value. Figure 4 As shown in the figure, HRRP images of three different types of aircraft (An-26, propeller-type aircraft, Yak-42) in the aircraft measurement data set are given. Figure 4(a) HRRP image measured in the HRRP target recognition experiment under the condition of class imbalance on the An-26 aircraft dataset. Figure 4 (b) is the HRRP image measured in the HRRP target recognition experiment under the condition of class imbalance on the propeller aircraft dataset. Figure 4 (c) is the HRRP image measured in the HRRP target recognition experiment under the condition of class imbalance on the Yak-42 aircraft dataset. Under the condition of class imbalance, the recognition accuracy of this method and other comparative recognition methods on the measured aircraft dataset is as follows: Figure 5 As shown in the figure, under the condition of class imbalance, the confusion matrix of the recognition accuracy of each category target of the method of the present invention on the measured aircraft data set is as follows: Figure 6 shown.

[0041] It should be understood that although Figure 1-Figure 2 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 1-Figure 2 At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0042] In one embodiment, Figure 7 As shown, a HRRP target recognition device based on an azimuth geometric constraint classifier is provided, comprising: a model building module 702, a feature learning module 704, a sub-prototype displacement acquisition module 706, an edge category feature training module 708 and a target image acquisition module 710, wherein: The model building module 702 is used to build a radar target recognition model based on HRRP. The radar target recognition model includes: a feature extraction module, an azimuth geometric constraint classifier and a loss function.

[0043] The feature learning module 704 is used to input the unbalanced HRRP data set into the feature extraction module to extract target features, and input the target features into the azimuth geometric constraint classifier for category learning. The target features include: dominant category features and marginal category features.

[0044] The sub-prototype displacement acquisition module 706 is used to train the dominant category feature to obtain the dominant category azimuth frame classification sub-prototype, and use a layer of convolution to calculate the sub-prototype displacement between the azimuth frame classification sub-prototype of the current dominant category feature and the next dominant category feature.

[0045] The edge category feature training module 708 is used to train the edge category feature according to the sub-prototype displacement to obtain the edge category azimuth frame classification sub-prototype.

[0046] The target image acquisition module 710 is used to reconstruct the loss function according to the dominant category azimuth frame classification sub-prototype and the marginal category azimuth frame classification sub-prototype to obtain the overall loss function, optimize the radar target recognition module according to the overall loss function, train the unbalanced HRRP data set through the optimized radar target recognition module, and output the target image.

[0047] For the specific limitations of the HRRP target recognition device based on the azimuth geometric constraint classifier, please refer to the limitations of the HRRP target recognition method based on the azimuth geometric constraint classifier above, which will not be repeated here. Each module in the above-mentioned HRRP target recognition device based on the azimuth geometric constraint classifier can be fully or partially implemented by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0048] Those skilled in the art will understand that Figure 7 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present invention, and does not constitute a limitation on the computer device to which the solution of the present invention is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0049] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided by the present invention can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0050] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0051] The above-mentioned embodiments only express several implementation modes of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the invention. It should be pointed out that, for a person of ordinary skill in the art, several modifications and improvements can be made without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the attached claims.

Claims

1. A HRRP target recognition method based on azimuth geometric constraint classifier, characterized in that: The method comprises: Constructing a radar target recognition model based on HRRP; the radar target recognition model includes: a feature extraction module, an azimuth geometric constraint classifier and a loss function; Inputting the unbalanced HRRP data set into the feature extraction module to extract target features, and inputting the target features into the azimuth geometric constraint classifier for category learning; the target features include: dominant category features and marginal category features; Training the dominant category feature to obtain a dominant category azimuth frame classification sub-prototype, and using a layer of convolution to calculate a sub-prototype displacement between the current dominant category feature and the azimuth frame classification sub-prototype of the next dominant category feature; The edge category feature is trained according to the sub-prototype displacement to obtain an edge category azimuth frame classification sub-prototype; The loss function is reconstructed according to the dominant category azimuth frame classification sub-prototype and the marginal category azimuth frame classification sub-prototype to obtain an overall loss function, the radar target recognition model is optimized according to the overall loss function, the unbalanced HRRP dataset is trained by the optimized radar target recognition model, and a target image is output.

2. The method according to claim 1, characterized in that Inputting the unbalanced HRRP data set into the feature extraction module to extract target features, and inputting the target features into the azimuth geometric constraint classifier for category learning, including: The unbalanced HRRP dataset is input into the feature extraction module, and the long short-term memory network is used to extract the change characteristics of the target strong scattering center in the unbalanced HRRP dataset to obtain the target features: ； in, is the target feature, For unbalanced HRRP data; Inputting the target features into the azimuth geometry classifier for category learning; The unbalanced HRRP dataset includes: dominant category data and marginal category data.

3. The method according to claim 2, characterized in that The dominant category feature is trained to obtain a dominant category azimuth frame classification sub-prototype, and a layer of convolution is used to calculate the sub-prototype displacement between the current dominant category feature and the azimuth frame classification sub-prototype of the next dominant category feature, including: The target azimuth frame classification sub-prototype corresponding to the target feature is learned by the azimuth geometric constraint classifier, and the dominant category feature is trained to obtain the dominant category azimuth frame classification sub-prototype, and the geometric constraint is defined according to the learned dominant category azimuth frame classification sub-prototype: ； in, is the set of sub-prototype displacements of the dominant category azimuth frame classification, is the number of learned target azimuth frame classification sub-prototypes, is the number of dominant categories; A layer of convolution is used to calculate the sub-prototype displacement between the azimuth frame classification sub-prototype of the current dominant category feature and the next dominant category feature: ； in, For the The leading category The dominant category azimuth frame classification sub-prototype and the The sub-prototype displacement between the azimuth frame classification sub-prototypes, For the transformation operation, For the The dominant category is learned through training dominant category azimuth frame classification sub-prototypes, For the The dominant category is learned through training Azimuth frame classification sub-prototypes.

4. The method according to claim 3, characterized in that The edge category feature is trained according to the sub-prototype displacement to obtain an edge category azimuth frame classification sub-prototype, including: The edge category features are trained by the azimuth geometric constraint classifier according to the sub-prototype displacement: ； in, For the The edge categories newly generated edge category azimuth frame classification sub-prototype, For the The edge categories are learned through the target features. edge category azimuth frame classification sub-prototypes, is the total number of learned edge category azimuth frame classification sub-prototypes, For the The leading category The dominant category azimuth frame classification sub-prototype and the The sub-prototype displacement between the azimuth frame classification sub-prototypes, For averaging operation, , .

5. The method according to claim 4, characterized in that After the step of training the edge category feature according to the sub-prototype displacement to obtain the edge category azimuth frame classification sub-prototype, the method further includes: The azimuth geometric constraint classifier uses cosine similarity as a distance metric to classify the target feature into the category of the classification sub-prototype that is closest to the current target feature: ； in, is the cosine similarity, is the number of target feature categories, is the total number of categories, is the total number of dominant categories, is the total number of edge categories, is the target feature, For the Category Azimuth frame classification sub-prototypes, is the distance measure.

6. The method according to claim 5, characterized in that Reconstructing the loss function according to the dominant category azimuth frame classification sub-prototype and the marginal category azimuth frame classification sub-prototype to obtain an overall loss function includes: The probability that the target feature belongs to the target azimuth frame classification sub-prototype is generated according to the dominant category azimuth frame classification sub-prototype, the marginal category azimuth frame classification sub-prototype and the current target feature: ； in, is the probability that the target feature belongs to the target azimuth frame classification sub-prototype, is the number of learned target azimuth frame classification sub-prototypes, For the Category Azimuth frame classification sub-prototypes; Reconstruct the classification loss generated by the azimuth set constraint classifier according to the probability: ； in, is the reconstructed classification loss function, is the HRRP data to be identified, The target category label identified by the model, For the Category Azimuth frame classification sub-prototypes; A decoupling loss function is constructed based on the distance metric and the dominant category azimuth frame classification sub-prototype: ； in, To decouple the loss function, is the distance metric, For the Category Azimuth frame classification sub-prototypes, is the distance metric The transpose of Construct an overall loss function based on the classification loss function and the decoupling loss function: ； in, is the overall loss function, For trade-off parameters; The radar target recognition model is optimized by the overall loss function, the unbalanced HRRP data set is trained by the optimized radar target recognition model, and a target image is output.

7. A HRRP target recognition device based on azimuth geometric constraint classifier, characterized in that: The device comprises: A model building module, used to build a radar target recognition model based on HRRP; the radar target recognition model includes: a feature extraction module, an azimuth geometric constraint classifier and a loss function; A feature learning module is used to input the unbalanced HRRP data set into the feature extraction module to extract target features, and input the target features into the azimuth geometric constraint classifier for category learning; the target features include: dominant category features and marginal category features; A sub-prototype displacement acquisition module is used to train the dominant category feature to obtain a dominant category azimuth frame classification sub-prototype, and use a layer of convolution to calculate the sub-prototype displacement between the current dominant category feature and the azimuth frame classification sub-prototype of the next dominant category feature; An edge category feature training module, used for training the edge category feature according to the sub-prototype displacement to obtain an edge category azimuth frame classification sub-prototype; The target image acquisition module is used to reconstruct the loss function according to the dominant category azimuth frame classification sub-prototype and the marginal category azimuth frame classification sub-prototype to obtain an overall loss function, optimize the radar target recognition module according to the overall loss function, train the unbalanced HRRP data set through the optimized radar target recognition module, and output a target image.

Images