Radar radiation source individual precision intelligent identification method, system, device and terminal

Abstract

The present application belongs to the technical field of individual identification of radar radiation source, and discloses a radar radiation source individual precision intelligent identification method, system, device and terminal, obtains the corresponding bispectrum of the received radar radiation source signal; extracts the features of the bispectrum according to the Laplace-Gaussian operator; inputs the extracted features into the deep residual network based on the norm and the dynamic learning rate for training to obtain the trained model; and realizes the individual intelligent identification of the radar radiation source signal by using the trained model. The radar radiation source individual precision intelligent identification method realizes the individual identification of the radiation source under the condition that the difference of the radar fingerprint characteristics is not obvious, improves the individual identification efficiency of the radiation source under the condition of ensuring the accuracy, improves the generalization, robustness and accuracy during use of the network model, effectively mines the difference characteristics between the individuals of the radiation source, and can achieve better identification effect under the condition that the radar fingerprint characteristics are similar.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of individual identification of radar radiation sources, and particularly relates to a radar radiation source individual accurate intelligent identification method, system, device and terminal. BACKGROUND

[0002] At present, radar radiation source identification, that is, after a series of analysis and processing of the received radar radiation source individual signal, it is determined that the received individual signal comes from which radar radiation source and its position and parameter information. Early radars are relatively simple, and radar radiation source identification is mostly through artificial processing and analysis of the inter-pulse parameters obtained by measuring the radar radiation source, and then compared and analyzed with the individual information of each radar radiation source in the known established database. However, in recent years, radar technology has developed rapidly, and various new radars have emerged in an endless stream, and the battlefield electromagnetic environment has become more complex, and the similarity of the fingerprint information between each radar radiation source is high, which makes it difficult to distinguish the radar signals intercepted by the radar receiving device. Therefore, finding a more accurate and fast identification method has become an urgent problem and main development direction in the field of radar radiation sources.

[0003] The individual identification technology of radiation sources based on fingerprint information has started in the last century. American scholar Boorstyn proposed the theory of "Specific Emitter Identification, SEI", which uses the intercepted specific radar radiation source signal to extract the fingerprint characteristics of the signal, compares and analyzes it with the previously established information library, classifies and identifies the radar radiation source that emits the signal according to the matching result, but this method relies on expert judgment, and human factors are large, which is easy to produce misjudgment. At the same time, the existing individual identification of radar radiation sources relies more on the difference between radar fingerprint information, and if the difference is small, it will seriously affect the identification result. Based on this, a method of using constellation diagram to extract fingerprint information is proposed abroad, and input into the convolutional neural network for identification; based on the statistics of radiation source conventional parameters such as direction of arrival (DOA), pulse width (PW), pulse repetition frequency (PRF) and radar frequency (RF), as the basis for classification and identification, the input network uses naive Bayes classifier, clustering, SVM and other methods; based on kernel principal component analysis (KPCA) prediction learning method; based on a method of using Frechet distance to calculate the distance between signals, pulse envelope or instantaneous frequency; based on bispectrum + SURF (Speed-up robust features) features. However, using part of the fingerprint features has high accuracy, but the calculation amount is huge, and the fingerprint features with small calculation amount often have low accuracy.

[0004] And for the domestic, there are based on the phase observation model + long short memory network method, based on VMD decomposition of radar radiation source individual accurate intelligent identification method, based on the method of constructing data set based on pulse data stream, based on the individual identification method of fuzzy function, based to the feature extraction of radar radiation source individual accurate intelligent identification method based on Hilbert transform, and based on the method of using deep belief network DBN and radiation source signal envelope characteristics for radiation source individual classification method. However, part of the network model is too fitted for the training set, resulting in high accuracy during training but unsatisfactory effect during test application. Therefore, it is urgent to find a fast and accurate radar fingerprint feature and highlight it, and to enlarge its difference; at the same time, improve the network model, reduce its dependence on the training set, and prevent overfitting.

[0005] Through the above analysis, the problems and defects of the prior art are:

[0006] (1) With the development of radar technology, the complexity of signal and the complexity of electromagnetic environment are increasing, which makes it difficult to meet the requirement of identification accuracy by inter-pulse parameters, and it is necessary to use intra-pulse modulation information. However, the existing radar radiation source individual identification using intra-pulse information is more dependent on the difference between radar radiation source fingerprint information. If the difference between the to-be-identified radiation source fingerprint information is small, the identification accuracy will be seriously reduced.

[0007] (2) The existing radiation source individual identification technology based on fingerprint information relies on expert judgment, and the identification error is large by using artificial method, so it is necessary to use artificial intelligence method to mine deeper abstract features for identification.

[0008] (3) When the existing radar radiation source individual identification method using artificial intelligence is used, the shallow network often cannot get good identification effect, and the deep network will greatly increase the time consumption, and there are problems of overfitting and network degradation. SUMMARY

[0009] In view of the problems existing in the prior art, the present application provides a radar radiation source individual accurate intelligent identification method, system, device and terminal, especially a radar radiation source individual accurate intelligent identification method, system, medium, device and terminal when the fingerprint features are similar.

[0010] The present application is realized in this way, a radar radiation source individual accurate intelligent identification method, the radar radiation source individual accurate intelligent identification method comprises: obtaining the bispectrum corresponding to the radar radiation source signal and performing Laplace-Gaussian operator feature extraction; constructing a trained model by using the extracted features, and realizing the accurate intelligent identification of radar radiation source individual by using the trained model.

[0011] Further, the radar radiation source individual accurate intelligent identification method comprises the following steps:

[0012] Step one, the corresponding bispectrum of the received radar emitter signal is obtained, and the deep features of the radar emitter signal are mined;

[0013] Step two, the bispectrum is extracted according to the Laplace-Gaussian operator, and the edge difference information between individual radar emitters is further highlighted;

[0014] Step three, the extracted features are input into the deep residual network based on the norm and dynamic learning rate for training, and the trained model is obtained, which achieves ideal recognition effect with less consumption;

[0015] Step four, the trained model is used to realize intelligent recognition of individual radar emitter signals.

[0016] Further, the step one of obtaining the corresponding bispectrum of the received radar emitter signal comprises:

[0017] The third-order cumulant of the radar emitter signal x(t) is calculated, and the expression is as follows:

[0018] C 3s (τ1,τ2)=E{s * (t)x(t+τ1)x(t+τ2)};

[0019] Wherein, x is the received signal, τ is the time delay, s * (t) is the conjugate signal; E is the mathematical expectation of the corresponding value, and the obtained C 3s (τ1,τ2) is the corresponding third-order cumulant.

[0020] The bispectrum of the third-order cumulant C 3s (τ1,τ2) is calculated, and the expression is as follows:

[0021]

[0022] Wherein, C 3s (τ1,τ2) is the corresponding third-order cumulant obtained, is the two-dimensional Fourier transform of the third-order cumulant, and the obtained B s (ω1,ω2) is the bispectrum.

[0023] Further, the step two of extracting the bispectrum according to the Laplace-Gaussian operator comprises:

[0024] The Laplace-Gaussian operator with 0 as the center and σ as the Gaussian standard deviation, the expression is as follows:

[0025]

[0026] Wherein, Gσ (x, y) is a second-order Gaussian function, and the expression is as follows:

[0027]

[0028] wherein, sigma is a Gaussian standard deviation.

[0029] Then, the Laplace-Gaussian operator feature extraction is performed on the bispectrum:

[0030] LoGB s (ω1, ω2) = LoG * B s (ω1, ω2);

[0031] wherein, * represents convolution operation, and the operator is used to perform convolution operation on the bispectrum.

[0032] Further, the step three of inputting the extracted features into the deep residual network based on norm and dynamic learning rate for training to obtain a trained model comprises:

[0033] When the network calculates the parameters of each layer, the L2 norm is introduced, and the square root of the sum of squares of each feature vector element is calculated to make the network more sparse and smooth.

[0034] ||x||2=(|x1| 2 +|x1| 2 +|x1| 2 +...+|x n | 2 ) 1 / 2 ;

[0035] In each round of learning, the dynamic learning rate is as follows:

[0036] lr(n) = lr * 0.2 [n / 10] ;

[0037] wherein, lr(n) is the current learning rate, n is the batch, lr is the preset learning rate, and the learning rate becomes 0.2 times of the current value after every 10 batch learning, so that the learning rate gradually decreases with the change of rounds.

[0038] The deep residual network is composed of various residual blocks, and the learning rule of each residual block is as follows:

[0039] F(x) = H(x) - x;

[0040] wherein, H(x) is the back propagation function, F(x) is the forward propagation function, and x is the network input.

[0041] The bispectrum extracted by the Laplace-Gaussian operator is input into a single channel to generate a feature of HxWx1, wherein H and W are the length and width of the feature map respectively, and 1 is the number of channels; the feature is input into a residual block 1 for convolution operation, and the result is input into a residual block 2 until the operation of the four residual blocks is completed; finally, the final recognition result is obtained through a pooling layer and a fully connected layer.

[0042] Further, the step four utilizes the trained model to realize individual intelligent identification of the radar radiation source signal.

[0043] The corresponding bispectrum is obtained by receiving the unclassified preprocessed signal of the trained radar radiation source individual, the bispectrum is characterized by the Laplace-Gaussian operator, and the trained network model is loaded; the feature is input into the network to obtain the recognition result; the probability matrix of each radar radiation source individual is obtained through the Softmax layer, and the one with the highest probability is the recognition result, and the Softmax expression is as follows:

[0044] M = max(z);

[0045]

[0046] Wherein, z is the result vector, max is the maximum value of z, and z i is the i-th result.

[0047] Another object of the present application is to provide a radar radiation source individual accurate intelligent identification system applying the radar radiation source individual accurate intelligent identification method.

[0048] A bispectrum obtaining module is configured to obtain the corresponding bispectrum of the received radar radiation source signal;

[0049] A feature extraction module is configured to extract the features of the bispectrum according to the Laplace-Gaussian operator;

[0050] A feature training module is configured to input the extracted features into a deep residual network based on the norm and the dynamic learning rate for training to obtain a trained model;

[0051] An individual identification module is configured to utilize the trained model to realize individual intelligent identification of the radar radiation source signal.

[0052] Another object of the present application is to provide a computer device, which comprises a memory and a processor, wherein the memory stores a computer program, and the computer program is executed by the processor to make the processor execute the steps of the radar radiation source individual accurate intelligent identification method.

[0053] Another object of the present application is to provide a computer readable storage medium storing a computer program which, when executed by a processor, causes the processor to perform the steps of the radar emitter individual accurate intelligent identification method.

[0054] Another object of the present application is to provide an information data processing terminal for implementing the radar emitter individual accurate intelligent identification system.

[0055] In combination with the above technical solutions and the technical problems solved, the technical solutions of the present application have the following advantages and positive effects:

[0056] First, in view of the technical problems existing in the prior art and the difficulty in solving the problems, the technical solutions of the present application are closely combined with the results and data obtained during the development process, and the technical problems solved by the technical solutions are analyzed in detail and deeply, and some creative technical effects brought about after the problems are solved are described as follows:

[0057] The radar emitter individual accurate intelligent identification method provided by the present application first obtains the corresponding bispectrum of the received radar emitter signal, and extracts features from the bispectrum based on the Laplacian of gaussain (LoG) operator; finally, the extracted features are input into a deep residual network (Resnet) based on a norm and a dynamic learning rate for training to obtain a trained model, and the model is used to realize individual intelligent identification of the radar emitter signal. The present application effectively mines the difference features between individual emitters, and can achieve better recognition effect in the case of similar radar fingerprint features.

[0058] The present application can effectively realize the identification of individual emitters in the case of similar radar fingerprint features, and has better performance than other methods. In addition, the radar emitter individual identification method provided by the present application is also applicable to the identification of individual emitters in the case of large difference in radar fingerprint features.

[0059] Second, from the perspective of the product as a whole, the technical solutions of the present application have the following technical effects and advantages:

[0060] The radar emitter individual accurate intelligent identification method provided by the present application realizes the identification of individual emitters in the case of insufficient radar fingerprint feature extraction and insignificant difference between each other in a complex electromagnetic environment, improves the efficiency of individual emitter identification while ensuring the accuracy of the network, and improves the generalization and robustness of the network model and the accuracy during use.

[0061] Third, as the invention of the claim of the creative auxiliary evidence, also embodied in the following several important aspects:

[0062] The radar radiation source individual precision intelligent identification method provided by the application combines neural networks, radar intra-pulse signal processing technology and pattern recognition theory.

[0063] (1) The commercial value of the technical scheme of the application after transformation is:

[0064] The radar radiation source individual identification technology of the application is truly used in actual radar radiation source individual signal identification, not just in theoretical simulation signal identification.

[0065] (2) The technical scheme of the application solves the technical problems that people have been eager to solve but have failed to succeed:

[0066] In today's complex electromagnetic environment, the radar radiation source features can be effectively extracted, the differences between radar radiation sources are amplified, and higher identification accuracy is provided when the radar radiation source fingerprints are similar. At the same time, the improved network prevents overfitting while ensuring training accuracy, does not need to stack network layers during training, reduces expenditure, and still maintains high accuracy during identification. BRIEF DESCRIPTION OF DRAWINGS

[0067] In order to more clearly illustrate the technical solutions of the embodiments of the application, the drawings needed in the embodiments of the application will be briefly introduced as follows. Obviously, the drawings described below are only some embodiments of the application, and other drawings can be obtained by those skilled in the art without creative labor on the basis of these drawings.

[0068] Figure 1 The radar radiation source individual precision intelligent identification method flowchart provided by the embodiments of the application;

[0069] Figure 2(a) is a schematic diagram of the radiation source individual identification accuracy and loss function using the radar radiation source individual precision intelligent identification method provided by the embodiments of the application;

[0070] Figure 2(b) is a schematic diagram of the radiation source individual identification accuracy and loss function using the bispectrum method provided by the embodiments of the application. DETAILED DESCRIPTION

[0071] In order to make the purpose, technical scheme and advantages of the application more clear and apparent, the application will be further described in detail below in combination with embodiments. It should be understood that the specific embodiments described herein are only used to explain the application and not to limit the application.

[0072] In view of the problems in the prior art, the present application provides a radar emitter individual accurate intelligent identification method, system, device and terminal, which will be described in detail below with reference to the drawings.

[0073] I. Explanation of embodiments. In order for those skilled in the art to fully understand how the present application is specifically implemented, this part is an explanation of the embodiments of the technical solutions of the claims.

[0074] As shown in Figure 1 , the radar emitter individual accurate intelligent identification method provided by the embodiments of the present application comprises the following steps:

[0075] S101, obtaining the corresponding bispectrum of the received radar emitter signal;

[0076] S102, performing feature extraction on the bispectrum according to the Laplacian of Gaussian operator;

[0077] S103, inputting the extracted features into a deep residual network based on norm and dynamic learning rate for training to obtain a trained model;

[0078] S104, realizing individual intelligent identification of the radar emitter signal by using the trained model.

[0079] As a preferred embodiment, the radar emitter individual accurate intelligent identification method provided by the embodiments of the present application when the fingerprint features are similar, specifically comprises the following steps:

[0080] Firstly, the corresponding bispectrum of the received radar emitter signal is obtained, and the bispectrum is subjected to feature extraction according to the Laplacian of Gaussian (LoG) operator, and the specific implementation process is as follows:

[0081] The third-order cumulant of the radar emitter signal x(t) is calculated, and the expression is as follows:

[0082] C 3s (τ1,τ2)=E{s * (t)x(t+τ1)x(t+τ2)}

[0083] Wherein, x is the received signal, τ is the time delay, s * (t) is its conjugate signal, E is the mathematical expectation of its corresponding value, and the obtained C 3s (τ1,τ2) is its corresponding third-order cumulant.

[0084] The bispectrum transform of the third-order cumulant C 3s (τ1,τ2) is calculated, and the expression is as follows:

[0085]

[0086] where C 3s (τ1, τ2) is the corresponding third-order cumulant to be obtained, is the two-dimensional Fourier transform of the third-order cumulant, and B s (ω1, ω2) is the bispectrum.

[0087] The Laplace-Gaussian operator expression centered at 0 with a Gaussian standard deviation of σ is as follows:

[0088]

[0089] where G σ (x, y) is the second-order Gaussian function, and is as follows:

[0090]

[0091] where σ is the Gaussian standard deviation.

[0092] Then, the Laplace-Gaussian operator feature extraction of the bispectrum is as follows:

[0093] LoGB s (ω1, ω2) = LoG * B s (ω1, ω2)

[0094] where * represents convolution operation, i.e., the operator is used for convolution operation on the bispectrum.

[0095] Secondly, the extracted features are input into a deep residual network (Resnet) based on the norm and dynamic learning rate for training to obtain a trained model, and the specific implementation process is as follows:

[0096] When the network calculates the parameters of each layer, the L2 norm is introduced, i.e., the square sum of each feature vector element is then squared to make the network more sparse and smooth.

[0097] ||x||2 = (|x1| 2 +|x1| 2 +|x1| 2 +...+|x n | 2 ) 1 / 2

[0098] The dynamic learning rate is used in each round of learning as follows:

[0099] lr(n) = lr * 0.2 [n / 10]

[0100] Wherein, lr(n) is a current learning rate, n is a batch, lr is a preset learning rate, that is, after 10 batches, the learning rate becomes the current 0.2 times, that is, the learning rate is gradually reduced with the change of rounds.

[0101] The deep residual network is composed of various residual blocks, and the learning rule of each residual block is:

[0102] F(x)=H(x)-x

[0103] Wherein, H(x) is a back propagation function, F(x) is a forward propagation function, and x is a network input.

[0104] The bispectrum extracted by the Laplacian of gaussain operator is input into a single channel to generate an HxWx1 feature, wherein H and W are the length and width of the feature map respectively, and 1 is the channel number, then the feature is input into residual block 1 for convolution operation, and the result is input into residual block 2, until after 4 residual block operations, finally through the pooling layer and the full connection layer, the final recognition result is obtained.

[0105] Thirdly, the model is used to realize individual intelligent identification of radar emitter signals, and the specific implementation process is:

[0106] The corresponding bispectrum of the received unclassified preprocessed signal of the trained radar emitter individual is obtained, the bispectrum is extracted according to the Laplacian of gaussain (LoG), and the trained network model is loaded. The feature is input into the network to obtain the recognition result, and the probability matrix of each radar emitter individual is obtained through the Softmax layer, and the highest probability is the recognition result, and the Softmax expression is as follows:

[0107] M=max(z)

[0108]

[0109] Wherein, z is a result vector, max is the maximum value of z, and z i Is the i-th result.

[0110] The radar emitter individual precise intelligent identification system provided by the embodiment of the application comprises:

[0111] A bispectrum obtaining module is configured to obtain the corresponding bispectrum of the received radar emitter signal;

[0112] A feature extraction module is configured to extract features from the bispectrum according to the Laplacian of gaussain operator;

[0113] The feature training module is configured to input the extracted features into a deep residual network based on a norm and a dynamic learning rate for training to obtain a trained model.

[0114] The individual identification module is configured to realize intelligent identification of radar emitter signal individuals by using the trained model.

[0115] II. Application Examples. In order to prove the creativity and technical value of the technical solutions of the present application, this part is an application example of the technical solutions of the claims on a specific product or related technology.

[0116] After intercepting the radar emitter individual signal, the signal is transmitted into the system after pre-processing of the receiver, the bispectrum of the signal is extracted by the method, and the Gaussian Laplace operator is used for further extraction of features to form a training set and a verification set. During training, the improved model is loaded, the trained model is obtained by using the training set, and the effect is verified by using the verification set. After intercepting the radar emitter individual signal, the signal is transmitted into the system after pre-processing of the receiver, the bispectrum of the signal is extracted by the method, and the Gaussian Laplace operator is used for further extraction of features, the trained model is used to determine which radar emitter individual transmits the signal, and the identification of the radar emitter individual is realized.

[0117] III. Evidence of the effects of the embodiments. The embodiments of the present application have achieved some positive effects during the research and development or use, and indeed have great advantages compared with the prior art. The following content is described in combination with the data and graphs of the test process.

[0118] The simulation experiment provided by the embodiment of the present application uses signals of five different radar radiation source individuals, the channel environment is Gaussian white noise, the signal-to-noise ratio is set to 10 dB, each individual has 1500 sample data for network training and 500 sample data for testing, so the total amount of training samples is 7500 and the total number of test set samples is 2500. During training, the control group extracts the bispectrum features of all signals in the sample set, inputs the deep residual network for training, adopts the SGD optimization method during the training process, the loss function is the cross-entropy loss function, the sample data amount of each batch during the training process is set to 32, and a total of 100 training batches are set. The present application extracts the bispectrum features of all signals in the sample set, further extracts the features by using the Laplace-Gaussian operator, inputs the improved deep residual network for training, adopts the SGD optimization method during the training process, the loss function is the cross-entropy loss function, the sample data amount of each batch during the training process is set to 32, and a total of 100 training batches are set. As shown in FIGS. 2(a)-(b), the horizontal axis is the training round, and the vertical axis is the accuracy and the loss function. It can be seen from the comparison of the training set and the validation set of the two methods that, on the training set and the validation set, the radar radiation source individual accurate intelligent identification method provided by the embodiment of the present application is compared with the bispectrum method, the accuracy is improved from 67% to 80%, and the loss function is greatly reduced, which proves the feasibility of the present application method under the condition that the radar fingerprint information is similar, and good results are achieved.

[0119] It should be noted that the embodiments of the present application can be realized by hardware, software or a combination of software and hardware. The hardware part can be realized by special logic; the software part can be stored in a memory and executed by a suitable instruction execution system, such as a microprocessor or a specially designed hardware. Those skilled in the art can understand that the above-mentioned devices and methods can be realized by computer executable instructions and / or included in processor control code, such as provided on a carrier medium, such as a magnetic disk, CD or DVD-ROM, a programmable memory, such as a read-only memory (firmware), or a data carrier, such as an optical or electronic signal carrier. The device of the present application and its modules can be realized by hardware circuit, such as ultra-large scale integrated circuit or gate array, semiconductor, such as logic chip, transistor, or programmable hardware device, such as field programmable gate array, programmable logic device, or software executed by various types of processors, or a combination of the above-mentioned hardware circuit and software, such as firmware.

[0120] The above is only a specific embodiment of the present application, but the protection scope of the present application is not limited thereto, any modification, equivalent replacement and improvement made by those skilled in the art within the technical range disclosed by the present application, within the spirit and principle of the present application, should be covered within the protection scope of the present application.

Claims

1. A radar emitter individual precision intelligent identification method, characterized in that, The radar radiation source individual precision intelligent identification method comprises the following steps: The radar radiation source individual precision intelligent identification method comprises the following steps: Step one, obtaining the corresponding bispectrum of the received radar radiation source signal; Step two, extracting features from the bispectrum according to the Laplace-Gaussian operator; Step three, inputting the extracted features into a deep residual network based on the norm and dynamic learning rate for training to obtain a trained model; Step four, using the trained model to realize individual intelligent identification of the radar radiation source signal; The step one comprises: Radar emitter signal The third order cumulant is calculated and expressed as follows: ； wherein is the received signal, is the time delay, is the conjugate signal; is the mathematical expectation of the corresponding value, the result is the corresponding third order cumulant; On the third order cumulants The bispectrum transform is calculated, and the expression is as follows: ； wherein, is the corresponding third order cumulant sought, is the two-dimensional Fourier transform of the third order cumulant sought, and is the bispectrum; The step two comprises: Centered on 0, The Laplace-Gaussian operator for the Gaussian standard deviation is expressed as follows: ； wherein is a second order Gaussian function, expressed as follows: ； wherein is the Gaussian standard deviation; Then the Laplace-Gaussian operator feature extraction is performed on the bispectrum: * ； Wherein, * represents convolution operation, and the operator is used for convolution operation on the bispectrum.

2. The radar emitter individual precision intelligent identification method of claim 1, wherein, The step three comprises: During network calculation of each layer parameter, L2 norm is introduced, and the square root of the sum of squares of each feature vector element is calculated to make the network more sparse and smooth; ； In each round of learning, a dynamic learning rate is used as follows: ； wherein, is the current learning rate, is the batch, is the preset learning rate, the learning rate becomes 0.2 times of the current one after every 10 batches, and the learning rate is gradually reduced as the round changes. The deep residual network is composed of various residual blocks, and the learning rule of each residual block is as follows: ； wherein, is a backpropagation function, is a forward propagation function, is a network input; The bispectrum extracted by the Laplace-Gaussian operator is input into a single channel to generate an HxWx1 feature, wherein H and W are the length and width of the feature map respectively, and 1 is the channel number; the feature is input into residual block 1 for convolution operation, and the result is input into residual block 2 until the operation of 4 residual blocks is completed; finally, the final identification result is obtained through the pooling layer and the fully connected layer.

3. The radar emitter individual precision intelligent identification method of claim 1, wherein, The step four comprises: In the step four, the corresponding bispectrum of the received radar radiation source signal is obtained, the features are extracted according to the Laplace-Gaussian operator, and the trained network model is loaded; the features are input into the network to obtain the identification result; the probability matrix of each radar radiation source individual is obtained through the Softmax layer, and the one with the highest probability is the identification result, and the Softmax expression is as follows: ； ； where z is the result vector, max is the maximum value of z, is the i-th result.

4. A radar emitter individual precision intelligent identification system applying the radar emitter individual precision intelligent identification method according to any one of claims 1-3, characterized in that, The radar radiation source individual precision intelligent identification system comprises: A bispectrum obtaining module for obtaining the corresponding bispectrum of the received radar radiation source signal; A feature extraction module for extracting features from the bispectrum according to the Laplace-Gaussian operator; A feature training module for inputting the extracted features into a deep residual network based on the norm and dynamic learning rate for training to obtain a trained model; An individual identification module for using the trained model to realize individual intelligent identification of the radar radiation source signal.

5. A computer device, comprising: The computer device comprises a memory and a processor, the memory stores a computer program, and the computer program is executed by the processor to make the processor execute the steps of the radar emitter individual precision intelligent identification method in any one of claims 1-3. 6.A computer readable storage medium, storing a computer program, the computer program is executed by a processor to make the processor execute the steps of the radar emitter individual precision intelligent identification method in any one of claims 1-3.

7. An information data processing terminal, characterized by The information data processing terminal is used to implement the radar emitter individual precision intelligent identification system in claim 4.

Images