Time-frequency-image fusion distance combined azimuth direction sar active jamming identification method and device

Abstract

The present application relates to a kind of time-frequency-image fusion distance combined azimuth SAR active jamming identification method and device, comprising: the radar jamming echo data to be identified is input into the interference identification network of training completion;Wherein, the radar jamming echo data to be identified includes the time-frequency spectrum of interference data to be identified and the SAR interference image to be identified;Interference identification network includes first feature extraction network, second feature extraction network, feature fusion network and classification network;First feature extraction network and second feature extraction network are connected to feature fusion network in parallel, and feature fusion network is connected to classification network;The interference type identification result corresponding to the radar jamming echo data to be identified is obtained.The present application can effectively extract and fuse the features from different modalities to achieve interference identification, with high recognition accuracy.

Description

Technical Field

[0001] This invention relates to the field of radar interference identification technology, specifically to a time-frequency-image fusion-based range-joint azimuth SAR active interference identification method and apparatus. Background Technology

[0002] In modern warfare, SAR (Synthetic Aperture Radar) serves as a crucial reconnaissance tool and is a significant research subject in the field of electronic jamming. To disrupt SAR reconnaissance, a series of jamming methods have been publicly reported in the literature, mainly categorized into active and passive jamming. Active jamming signals, in particular, pose significant challenges to the normal and effective operation of SAR. To effectively address the challenges of SAR jamming technology, many scholars have proposed a series of SAR jamming suppression and anti-jamming techniques. From the perspective of SAR reconnaissance and intelligence generation, the identification of SAR jamming types is a prerequisite and foundation for applying SAR anti-jamming techniques, and thus has extremely important research significance. Existing radar jamming identification methods primarily focus on identifying active jamming in the range direction, while research on identifying active jamming types that consider both the range and azimuth directions is relatively limited. Summary of the Invention

[0003] To address the aforementioned problems in the prior art, this invention provides a time-frequency-image fusion-based range-joint azimuth SAR active interference identification method and apparatus.

[0004] According to a first aspect of the present invention, a method for identifying active interference in SAR using time-frequency-image fusion is provided, the method comprising:

[0005] The radar interference echo data to be identified is input into the trained interference identification network; wherein, the radar interference echo data to be identified includes the time spectrum of the interference data to be identified and the SAR interference image to be identified; the interference identification network includes a first feature extraction network, a second feature extraction network, a feature fusion network, and a classification network; the first feature extraction network and the second feature extraction network are connected in parallel to the feature fusion network, and the feature fusion network is connected to the classification network; the first feature extraction network includes a convolutional module, and the second feature extraction network includes multiple alternately connected CBR layers and WavePool layers; the feature fusion network includes a cross-attention module and a global average pooling layer connected in sequence;

[0006] The interference type identification result corresponding to the radar interference echo data to be identified is obtained.

[0007] Optionally, obtaining the interference type identification result corresponding to the radar interference echo data to be identified includes:

[0008] The first feature is obtained by processing the time spectrum of the interference data to be identified based on the first feature extraction network;

[0009] The second feature is obtained by processing the SAR interference image to be identified using the second feature extraction network.

[0010] The first feature and the second feature are fused according to the feature fusion network and input into the classification network to obtain the interference type identification result corresponding to the radar interference echo data to be identified.

[0011] Optionally, the convolutional module includes a BCL layer, a first CBL layer, a second CBL layer, a CB layer, and a first CBR layer connected in sequence.

[0012] Optionally, the BCL layer includes a BN layer, a Conv layer, and a ReLU6 activation function connected in sequence; the first CBL layer and the second CBL layer have the same structure, both including a Conv layer, a BN layer, and a ReLU6 activation function connected in sequence; the CB layer includes a Conv layer and a BN layer connected in sequence; the first CBR layer includes a Conv layer, a BN layer, and a ReLU activation function connected in sequence.

[0013] Optionally, the second feature extraction network includes a second CBR layer, a first WavePool layer, a third CBR layer, a second WavePool layer, and a fourth CBR layer connected in sequence.

[0014] Optionally, the second CBR layer, the third CBR layer, and the fourth CBR layer have the same structure, each including a Conv layer, a BN layer, and a ReLU activation function connected in sequence.

[0015] Optionally, the training process of the interference identification network includes:

[0016] Construct a training dataset; wherein the training dataset includes a time-spectrum dataset of interference data and a SAR interference image dataset;

[0017] The interference detection network is trained using the training dataset, and the parameters of the interference detection network are adjusted according to a pre-constructed loss function.

[0018] Continue training the interference recognition network with adjusted parameters until the maximum number of iterations is reached, resulting in a trained interference loss network.

[0019] Optionally, the pre-constructed loss function is represented as follows:

[0020]

[0021] Among them, y i P represents the true label of the i-th sample in the training dataset. ic Let N represent the probability that the interference identification network predicts the category of the i-th sample as c, where N is the total number of samples and K is the number of categories of the samples.

[0022] According to a second aspect of the present invention, a time-frequency-image fusion range-joint azimuth SAR active interference identification device is provided, the device comprising:

[0023] An identification module is used to input radar interference echo data to be identified into a trained interference identification network; wherein, the radar interference echo data to be identified includes the time spectrum of the interference data to be identified and the SAR interference image to be identified; the interference identification network includes a first feature extraction network, a second feature extraction network, a feature fusion network, and a classification network; the first feature extraction network and the second feature extraction network are connected in parallel to the feature fusion network, and the feature fusion network is connected to the classification network; the first feature extraction network includes a convolutional module, the second feature extraction network includes multiple alternately connected CBR layers and WavePool layers; the feature fusion network includes a cross-attention module and a global average pooling layer connected in sequence;

[0024] The result acquisition module is used to obtain the interference type identification result corresponding to the radar interference echo data to be identified.

[0025] The technical solution provided by this invention may include the following beneficial effects:

[0026] Through the above technical solutions, the interference identification network constructed by this invention can extract and fuse features from different modalities of radar interference data, avoiding the shortcomings of manually extracting radar interference data features and achieving effective extraction of radar interference data features. Furthermore, this invention extracts features from both the time spectrum and SAR image modes of radar interference data, using a cross-attention module for complementary fusion of features between modes, giving different attention to feature maps of different modes and locations, providing more information for subsequent identification, and solving the problem that existing methods can only identify single-range SAR interference and have low generalization. This invention has the advantages of simultaneously identifying range and azimuth interference with high accuracy. Moreover, multiple convolutional layers are added to the feature extraction network to enhance the representation features of the neural network. In summary, this invention can effectively extract and fuse features from different modalities and achieve interference identification.

[0027] Other features and advantages of the present invention will be described in detail in the following detailed description section. Attached Figure Description

[0028] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the following detailed description to explain the invention, but do not constitute a limitation thereof. In the drawings:

[0029] Figure 1 This is a flowchart illustrating a time-frequency-image fusion-based range-joint azimuth SAR active interference identification method according to an exemplary embodiment.

[0030] Figure 2 This is a schematic diagram illustrating the structure of an interference identification network according to an exemplary embodiment.

[0031] Figure 3 This is a block diagram illustrating a time-frequency-image fusion range-joint azimuth SAR active interference identification device according to an exemplary embodiment. Detailed Implementation

[0032] Figure 1 This is a flowchart illustrating a time-frequency-image fusion range-joint azimuth SAR active interference identification method according to an exemplary embodiment, such as... Figure 1 As shown, the method includes the following steps:

[0033] S101. Input the radar interference echo data to be identified into the trained interference identification network; wherein, the radar interference echo data to be identified includes the time spectrum of the interference data to be identified and the SAR interference image to be identified; the interference identification network includes a first feature extraction network, a second feature extraction network, a feature fusion network, and a classification network; the first feature extraction network and the second feature extraction network are connected in parallel to the feature fusion network, and the feature fusion network is connected to the classification network; the first feature extraction network includes a convolutional module, the second feature extraction network includes multiple alternately connected CBR layers and WavePool layers; the feature fusion network includes a cross-attention module and a global average pooling layer connected in sequence.

[0034] Optionally, Figure 2 This is a schematic diagram illustrating the structure of an interference identification network according to an exemplary embodiment, such as... Figure 2 As shown, the convolutional module includes a BCL layer, a first CBL layer, a second CBL layer, a CB layer, and a first CBR layer connected in sequence.

[0035] It is understood that the BCL layer includes a BN layer, a Conv layer, and a ReLU6 activation function connected in sequence; the first CBL layer and the second CBL layer have the same structure, both including a Conv layer, a BN layer, and a ReLU6 activation function connected in sequence; the CB layer includes a Conv layer and a BN layer connected in sequence; and the first CBR layer includes a Conv layer, a BN layer, and a ReLU activation function connected in sequence.

[0036] It is worth mentioning that the BCL layer, the first CBL layer, and the second CBL layer use the ReLU6 activation function instead of the activation function used in conventional convolutional neural network feature extraction; limiting the output range reduces the dynamic range of network parameters, which helps to make the model lightweight; the first CBR layer uses 1×1 convolution to recombine the feature maps output by the depth convolution to achieve feature enhancement and realize the range-oriented feature extraction of SAR active interference.

[0037] Optionally, refer to Figure 2 The second feature extraction network includes a second CBR layer, a first WavePool layer, a third CBR layer, a second WavePool layer, and a fourth CBR layer connected in sequence.

[0038] Understandably, the second, third, and fourth CBR layers have the same structure, all including sequentially connected Conv layers, BN layers, and ReLU activation functions. Using WavePool layers instead of the pooling layers in conventional convolutional neural networks, and leveraging wavelet transform principles for multi-scale analysis, allows for the capture of feature information at different frequencies. This is better suited for complex scenarios like SAR images, improving the effectiveness of feature extraction. The fourth CBR layer also uses 1×1 convolutions to recombine the feature maps output by deep convolutions to achieve feature enhancement, enabling the extraction of range and azimuth features from SAR active interference.

[0039] It is worth mentioning that the feature fusion network includes a cross-attention module and a global average pooling layer. The cross-attention module significantly improves the model's representation ability and enhances the complementarity between information by dynamically allocating weights for features of different modalities. At the same time, it adaptively adjusts the importance of features, effectively reducing background noise interference and minimizing the loss of key information during the fusion process, thereby improving the model's generalization ability when dealing with complex scenes. The classification network includes a first connected layer and a second fully connected layer.

[0040] In one implementation, reference Figure 2Using the basic structure of a convolutional neural network, an interference recognition network is built. Based on the enhanced time-frequency map feature vector f1 output by the first feature extraction network, a key vector K and a value vector V are generated. The SAR image feature vector f2 output by the second feature extraction network is used to generate a query vector Q. The feature vector of the cross-attention module is A = MHA(Q,K,V), where A represents the cross-modal attention between the time-frequency branch and the SAR image branch. After computation, a global average pooling layer is used for downsampling, followed by a first fully connected layer (FC). This FC receives the feature map from the previous layer and converts it into a one-dimensional feature vector. The converted one-dimensional feature vector is then output to the second fully connected layer, which is directly connected to the output layer. The number of neurons in the second fully connected layer corresponds to the number of categories in the classification task.

[0041] In one implementation, the parameters of each network layer in the interference identification network can be referenced in Table 1: Detailed parameters of each network layer are shown in Table 1:

[0042] Table 1

[0043]

[0044]

[0045] Optionally, the training process of the interference identification network includes:

[0046] Construct a training dataset; the training dataset includes a time-spectrum dataset of interference data and a SAR interference image dataset;

[0047] The interference detection network is trained using the training dataset, and its parameters are adjusted according to a pre-built loss function.

[0048] Continue training the interference recognition network with adjusted parameters until the maximum number of iterations is reached, resulting in a trained interference loss network.

[0049] Optionally, the pre-constructed loss function is represented as follows:

[0050]

[0051] Among them, y i Let P represent the true label of the i-th sample in the training dataset. ic Let N represent the probability that the interference identification network predicts the class of the i-th sample as c, where N is the total number of samples and K is the number of classes of the samples.

[0052] Understandably, the time-spectrum dataset of interference data in the training dataset can be constructed as follows: Set the short-time Fourier transform parameters: window function is Hamming window, window length is 65, step size is 1, and number of Fourier transform points is 128; perform short-time Fourier transform on the simulated interference data to obtain the time-spectrum dataset of interference data;

[0053] The SAR interference image dataset in the training dataset can be constructed as follows: Set the SAR system parameters: carrier frequency 18GHz, bandwidth 300MHz, pulse width 1us, pulse repetition frequency 2000Hz; perform imaging processing on the simulated interference echo data to obtain the SAR interference image dataset.

[0054] After constructing the training dataset, preprocessing is required. The maximum value normalization method can be used to normalize the time-spectrum dataset of interference data and the SAR interference image dataset, respectively.

[0055] S102. Obtain the interference type identification result corresponding to the radar interference echo data to be identified.

[0056] Optionally, the interference type identification result corresponding to the radar interference echo data to be identified is obtained, including:

[0057] The first feature is obtained by processing the spectrum of the interference data to be identified using the first feature extraction network;

[0058] The second feature is obtained by processing the SAR interference image to be identified using the second feature extraction network;

[0059] The first and second features are fused using a feature fusion network and then input into a classification network to obtain the interference type identification result corresponding to the radar interference echo data to be identified.

[0060] Understandably, after fusing the first and second features using the feature fusion network, the interference type identification result can be obtained using the classification network and then processed through the Softmax layer. Specifically, the radar interference echo data to be identified is input into the trained interference identification network for forward propagation. After passing through two fully connected layers and the Softmax layer, the final interference type identification result is obtained. The identification results are shown in Table 2. The Softmax layer operation method is as follows:

[0061]

[0062] Among them, z j Let A(z) represent the output value of the j-th category, where j (j = 1, 2, ..., K) represents the number of categories. j) represents the probability of the j-th type of interference, with a value range of (0,1). The category with the highest probability is taken as the predicted interference category.

[0063] Table 2

[0064]

[0065] Through the above technical solutions, the interference identification network constructed in this invention can perform feature extraction and fusion of different modal data of radar interference. When extracting time-frequency features, the ReLU6 activation function is used instead of the activation function used in conventional convolutional neural network feature extraction; the output range is limited, which reduces the dynamic range of network parameters and helps to make the model lightweight; and a 1×1 convolution is added at the end to recombine the feature map output by the deep convolution to achieve feature enhancement effect, realizing the range feature extraction of SAR active interference; when extracting SAR image features, the WavePool layer is used instead of the pooling layer in the conventional convolutional neural network. By using the wavelet transform principle and multi-scale analysis, feature information at different frequencies can be captured, which is suitable for complex scenarios such as SAR images, improves the effectiveness of feature extraction, and realizes the range and azimuth feature extraction of SAR active interference. Finally, a cross-attention module is used for complementary fusion of features between modalities. This module significantly improves the model's representational ability and enhances the complementarity of information by dynamically allocating the weights of features from different modalities. Simultaneously, it adaptively adjusts feature importance, effectively reducing background noise interference and minimizing the loss of key information during fusion, thus improving the model's generalization ability in complex scenarios. The entire process avoids the drawbacks of manually extracting radar interference data features, achieving effective extraction of radar interference data features. Furthermore, this invention extracts features from both the time spectrum and SAR image modalities of radar interference data, using the cross-attention module for complementary fusion of features between modalities. Different attention is given to feature maps of different modalities and locations, providing more information for subsequent identification. This solves the problem that existing methods can only identify single-range SAR interference and have low generalization, giving this invention the advantages of simultaneously identifying range and azimuth interference with high accuracy. In summary, this invention can effectively extract and fuse features from different modalities and achieve interference identification.

[0066] Figure 3 This is a block diagram illustrating a time-frequency-image fusion range-joint azimuth SAR active interference identification device according to an exemplary embodiment, such as... Figure 3 As shown, the device 300 includes:

[0067] The identification module 301 is used to input the radar interference echo data to be identified into the trained interference identification network; wherein, the radar interference echo data to be identified includes the time spectrum of the interference data to be identified and the SAR interference image to be identified; the interference identification network includes a first feature extraction network, a second feature extraction network, a feature fusion network, and a classification network; the first feature extraction network and the second feature extraction network are connected in parallel to the feature fusion network, and the feature fusion network is connected to the classification network; the first feature extraction network includes a convolutional module, the second feature extraction network includes multiple alternately connected CBR layers and WavePool layers; the feature fusion network includes a cross-attention module and a global average pooling layer connected in sequence;

[0068] The result acquisition module 302 is used to obtain the interference type identification result corresponding to the radar interference echo data to be identified.

[0069] Through the above technical solutions, the interference identification network constructed by this invention can extract and fuse features from different modalities of radar interference data, avoiding the shortcomings of manually extracting radar interference data features and achieving effective extraction of radar interference data features. Furthermore, this invention extracts features from both the time spectrum and SAR image modes of radar interference data, using a cross-attention module for complementary fusion of features between modes, giving different attention to feature maps of different modes and locations, providing more information for subsequent identification, and solving the problem that existing methods can only identify single-range SAR interference and have low generalization. This invention has the advantages of simultaneously identifying range and azimuth interference with high accuracy. Moreover, multiple convolutional layers are added to the feature extraction network to enhance the representation features of the neural network. In summary, this invention can effectively extract and fuse features from different modalities and achieve interference identification.

[0070] Regarding the apparatus in the above embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0071] The preferred embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited to the specific details of the above embodiments. Within the scope of the technical concept of the present invention, various simple modifications can be made to the technical solution of the present invention, and these simple modifications all fall within the protection scope of the present invention.

[0072] It should also be noted that the various specific technical features described in the above specific embodiments can be combined in any suitable manner without contradiction. In order to avoid unnecessary repetition, the present invention will not describe the various possible combinations separately.

[0073] Furthermore, various different embodiments of the present invention can be combined in any way, as long as they do not violate the spirit of the present invention, they should also be regarded as the content disclosed by the present invention.

Claims

1. A time-frequency-image fusion method for identifying range-joint azimuth SAR active interference, characterized in that, The method includes: The radar interference echo data to be identified is input into the trained interference identification network; wherein, the radar interference echo data to be identified includes the time spectrum of the interference data to be identified and the SAR interference image to be identified; the interference identification network includes a first feature extraction network, a second feature extraction network, a feature fusion network, and a classification network; the first feature extraction network and the second feature extraction network are connected in parallel to the feature fusion network, and the feature fusion network is connected to the classification network; the first feature extraction network includes a convolutional module, and the second feature extraction network includes multiple alternately connected CBR layers and WavePool layers; the feature fusion network includes a cross-attention module and a global average pooling layer connected in sequence; The interference type identification result corresponding to the radar interference echo data to be identified is obtained.

2. The time-frequency-image fusion range-joint azimuth SAR active interference identification method according to claim 1, characterized in that, The process of obtaining the interference type identification result corresponding to the radar interference echo data to be identified includes: The first feature is obtained by processing the time spectrum of the interference data to be identified based on the first feature extraction network; The second feature is obtained by processing the SAR interference image to be identified using the second feature extraction network. The first feature and the second feature are fused according to the feature fusion network and input into the classification network to obtain the interference type identification result corresponding to the radar interference echo data to be identified.

3. The time-frequency-image fusion range-joint azimuth SAR active interference identification method according to claim 1, characterized in that, The convolutional module includes a BCL layer, a first CBL layer, a second CBL layer, a CB layer, and a first CBR layer connected in sequence.

4. The time-frequency-image fusion range-joint azimuth SAR active interference identification method according to claim 3, characterized in that, The BCL layer includes a BN layer, a Conv layer, and a ReLU6 activation function connected in sequence; the first CBL layer and the second CBL layer have the same structure, both including a Conv layer, a BN layer, and a ReLU6 activation function connected in sequence; the CB layer includes a Conv layer and a BN layer connected in sequence; the first CBR layer includes a Conv layer, a BN layer, and a ReLU activation function connected in sequence.

5. The time-frequency-image fusion range-joint azimuth SAR active interference identification method according to claim 1, characterized in that, The second feature extraction network includes a second CBR layer, a first WavePool layer, a third CBR layer, a second WavePool layer, and a fourth CBR layer connected in sequence.

6. The time-frequency-image fusion range-joint azimuth SAR active interference identification method according to claim 5, characterized in that, The second, third, and fourth CBR layers have the same structure, each including a Conv layer, a BN layer, and a ReLU activation function connected in sequence.

7. The time-frequency-image fusion range-joint azimuth SAR active interference identification method according to claim 1, characterized in that, The training process of the interference identification network includes: Construct a training dataset; wherein the training dataset includes a time-spectrum dataset of interference data and a SAR interference image dataset; The interference detection network is trained using the training dataset, and the parameters of the interference detection network are adjusted according to a pre-constructed loss function. Continue training the interference recognition network with adjusted parameters until the maximum number of iterations is reached, resulting in a trained interference loss network.

8. The time-frequency-image fusion range-joint azimuth SAR active interference identification method according to claim 7, characterized in that, The pre-constructed loss function is expressed as follows: Among them, y i P represents the true label of the i-th sample in the training dataset. ic Let N represent the probability that the interference identification network predicts the category of the i-th sample as c, where N is the total number of samples and K is the number of categories of the samples.

9. A time-frequency-image fusion range-joint azimuth SAR active interference identification device, characterized in that, The device includes: An identification module is used to input radar interference echo data to be identified into a trained interference identification network; wherein, the radar interference echo data to be identified includes the time spectrum of the interference data to be identified and the SAR interference image to be identified; the interference identification network includes a first feature extraction network, a second feature extraction network, a feature fusion network, and a classification network; the first feature extraction network and the second feature extraction network are connected in parallel to the feature fusion network, and the feature fusion network is connected to the classification network; the first feature extraction network includes a convolutional module, the second feature extraction network includes multiple alternately connected CBR layers and WavePool layers; the feature fusion network includes a cross-attention module and a global average pooling layer connected in sequence; The result acquisition module is used to obtain the interference type identification result corresponding to the radar interference echo data to be identified.

Images