An electromagnetic interference identification method, device, equipment and readable storage medium

Abstract

The application discloses an electromagnetic interference identification method and device, equipment and a readable storage medium, and relates to the technical field of signal processing. The in-phase and quadrature data of a to-be-identified radar signal in multiple distance banks is acquired, and the maximum power variation coefficient, the skewness coefficient, the kurtosis coefficient and the maximum sharpness of the to-be-identified radar signal are calculated according to the in-phase and quadrature data. The above characteristic parameters are input into a pre-trained electromagnetic interference identification model, the electromagnetic interference identification model adopts a full connection neural network structure, the identification results of the to-be-identified radar signal in the multiple distance banks are obtained through an output layer, and the identification results are used for distinguishing meteorological echo useful signals from electromagnetic interference signals. The pollution area of the to-be-identified radar signal is determined according to the identification results of the to-be-identified radar signal in the multiple distance banks, the pollution area is recorded, and an alarm prompt is sent to technical personnel, so that the missed judgment of electromagnetic interference and the misjudgment of useful signals can be avoided, and the identification accuracy of various electromagnetic interferences is improved.

Description

[0001] This application is a divisional application of a patent application entitled “An Electromagnetic Interference Identification Method, Apparatus, Device and Readable Storage Medium”, the original application being filed on April 24, 2023, application number 202310446527.6. Technical Field

[0002] This application relates to the field of signal processing technology, and more specifically, to an electromagnetic interference identification method, apparatus, device, and readable storage medium. Background Technology

[0003] Weather radar is a crucial facility for meteorological monitoring and severe weather warning. During operation, weather radar is frequently subjected to electromagnetic interference (EMI), which severely impacts its normal functioning. EMI refers to interference caused when the carrier wave of an interfering signal is the same as or close to the radar's transmission frequency, and this interference signal enters the radar receiver through the radar antenna. To improve the quality of the radar signal received by weather radar, it is necessary to detect whether the radar signal is subject to EMI. Existing solutions use radial noise algorithms and SQI (Speech Quality Index) algorithms to detect EMI. However, the radial noise algorithm can only identify spoke-shaped EMI and cannot identify helical EMI, resulting in low accuracy. While the SQI algorithm has some effectiveness in identifying both spoke-shaped and helical EMI, it is prone to missing EMI detection and misidentifying useful signals, leading to low accuracy. Summary of the Invention

[0004] In response, this application provides an electromagnetic interference identification method, apparatus, device, and readable storage medium to avoid missing electromagnetic interference and misjudging useful signals, thereby improving the accuracy of identifying various types of electromagnetic interference.

[0005] To achieve the above objectives, this application provides the following technical solution.

[0006] This application provides an electromagnetic interference identification method applied to a scenario of identifying co-frequency asynchronous interference in weather radar. The co-frequency asynchronous interference manifests as spokes, spirals, or speckles on the radar image. The electromagnetic interference identification method includes: Acquire basic data of the radar signal to be identified, including in-phase orthogonal data of the radar signal to be identified in multiple range databases; The maximum power variation coefficient, skewness coefficient, kurtosis coefficient, and maximum sharpness of the radar signal to be identified are calculated based on the in-phase orthogonal data of the radar signal to be identified in multiple range databases. The maximum power variation coefficient, skewness coefficient, kurtosis coefficient and maximum sharpness of the radar signal to be identified are input into a pre-trained electromagnetic interference identification model. The electromagnetic interference identification model adopts a fully connected neural network structure, and the identification results of the radar signal to be identified in multiple range databases are obtained through the output layer. The identification results are used to distinguish between useful meteorological echo signals and electromagnetic interference signals. Based on the identification results of the radar signal to be identified in multiple range databases, the contaminated area of ​​the radar signal to be identified is determined, the contaminated area is recorded, and an alarm is issued to the technicians.

[0007] Optionally, the step of calculating the maximum power variation coefficient of the radar signal to be identified based on in-phase orthogonal data of the radar signal to be identified in multiple range databases specifically includes: The power data of the radar signal to be identified is calculated using the formula: X(k) i =x(k) i conj(x(k) i ); where x(k) i This represents the in-phase orthogonal data of the i-th radar signal in the k-th range database; conj() is the conjugate operation; X(k) i This represents the power data of the i-th radar signal in the k-th range database; The maximum power variation coefficient Maxv is calculated using the formula: MaxV(k) = Max(X) - Min(X); where X is X(k). i The abbreviation for ; MaxV(k) is the maximum power change coefficient of the k-th distance library.

[0008] Optionally, the step of calculating the skewness coefficient and kurtosis coefficient of the radar signal to be identified based on in-phase orthogonal data of the radar signal to be identified in multiple range databases specifically includes: The skewness coefficient Sk is calculated using the following formula: Where E is the expected value of the in-phase orthogonal data in all distance databases; σ is the sample standard deviation, μ is the mean; and Sk(k) is the skewness coefficient of the k-th distance database. The kurtosis coefficient Ku is calculated using the following formula: Where, Ku(k) is the kurtosis coefficient of the k-th distance library.

[0009] Optionally, the step of calculating the maximum sharpness of the radar signal to be identified based on in-phase orthogonal data of the radar signal to be identified in multiple range databases specifically includes: Define the operator Lap(n). The typical operator for n=3 is as follows: ; The maximum sharpness is calculated based on the operator Lap(n), and the formula is: Macu(k) = max(conv(X) k ,Lap)); where X k =10 log 10 (abs(x(k) i )), i = 0, 1, ... M-1; conv represents convolution; abs represents absolute value; Macu(k) is the maximum sharpness of the k-th distance library.

[0010] Optionally, the electromagnetic interference identification model adopts a fully connected neural network structure; the fully connected neural network structure has N hidden layers, and the first N-1 layers are constructed using fully connected layers with tanh as the activation function; the Dropout algorithm is used in the fully connected neural network structure to prevent some neurons from being activated during training; the output layer of the fully connected neural network structure does not include an activation function, and since it is necessary to distinguish between electromagnetic interference and non-electromagnetic interference, the number of output parameters is 2; softmax is used to convert the output results into probabilities for the calculation of the loss function.

[0011] Optionally, during the training process of the electromagnetic interference identification model, the loss function is cross-entropy, the optimizer is Adam optimizer, and the backpropagation algorithm is used to update the parameter weights W and bias b; during the training process, the initial learning efficiency is set to a constant, the initial weights are set to an all-zero matrix, and the initial bias is set to an all-zero vector.

[0012] Optionally, the electromagnetic interference identification model adopts a fully connected neural network structure, and obtains the identification results of the radar signal to be identified in multiple range databases through the output layer, specifically including: The maximum power variation coefficient, skewness coefficient, kurtosis coefficient, and maximum sharpness of the radar signal to be identified in multiple range databases are input into the electromagnetic interference identification model. The electromagnetic interference identification model outputs the identification result of the radar signal to be identified in each range database to determine whether the radar signal to be identified is subject to electromagnetic interference in each range database.

[0013] This application embodiment also provides an electromagnetic interference identification device, applied to the scenario of identifying co-frequency asynchronous interference in weather radar. The co-frequency asynchronous interference appears as spokes, spirals, or speckles on the radar image. The electromagnetic interference identification device includes: An acquisition unit is used to acquire basic data of the radar signal to be identified, the basic data including in-phase orthogonal data of the radar signal to be identified in multiple range databases; The calculation unit is used to calculate the maximum power variation coefficient, skewness coefficient, kurtosis coefficient, and maximum sharpness of the radar signal to be identified based on the in-phase orthogonal data of the radar signal to be identified in multiple range databases. The determination unit is used to input the maximum power variation coefficient, skewness coefficient, kurtosis coefficient and maximum sharpness of the radar signal to be identified into a pre-trained electromagnetic interference identification model. The electromagnetic interference identification model adopts a fully connected neural network structure and obtains the identification results of the radar signal to be identified in multiple range databases through the output layer. The identification results are used to distinguish between useful meteorological echo signals and electromagnetic interference signals. The determining unit is further configured to determine the contaminated area of ​​the radar signal to be identified based on the identification results of the radar signal to be identified in multiple distance databases, record the contaminated area, and issue an alarm prompt to the technicians.

[0014] This application also provides a computer device, including: a memory, a processor, and a bus system; wherein, the memory is used to store a program; the bus system is used to connect the memory and the processor to enable communication between the memory and the processor; the processor is used to execute the program in the memory to implement the electromagnetic interference identification method.

[0015] This application also provides a computer-readable storage medium storing instructions that, when run on a computer, cause the computer to execute the electromagnetic interference identification method.

[0016] Compared with the prior art, this application has at least the following beneficial effects: This application calculates the characteristic parameters (including maximum power variation coefficient, skewness coefficient, kurtosis coefficient, and maximum sharpness) of the radar signal to be identified based on the basic data of the radar signal to be identified (including in-phase orthogonal data of the radar signal to be identified in multiple range databases), and inputs them into a pre-trained electromagnetic interference identification model. This can achieve accurate identification of various spoke-shaped, spiral, or speckled electromagnetic interferences, and can avoid missed detection of electromagnetic interference and misjudgment of useful signals, greatly improving the accuracy of radar signal identification of whether it is subject to electromagnetic interference. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0018] Figure 1A flowchart illustrating an electromagnetic interference identification method provided in an embodiment of this application; Figure 2 This is a schematic diagram of the structure of a neural network model provided in an embodiment of this application; Figure 3 This is a schematic diagram of an electromagnetic interference identification device provided in an embodiment of this application. Detailed Implementation

[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a particular order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0021] Weather radar is a crucial facility for meteorological monitoring and early warning of severe weather. However, weather radar is frequently subjected to electromagnetic interference (EMI) during operation, severely impacting its normal functioning. EMI refers to interference caused when the carrier wave of an interfering signal is the same as or close to the radar's transmission frequency, entering the radar receiver through the radar antenna. EMI is divided into synchronous interference and asynchronous interference. Synchronous interference typically appears as concentric circles on the radar image, while asynchronous interference typically appears as a spiral. Asynchronous interference from a non-rotating transmitting antenna source appears as spokes on the radar image. In Chinese weather radar, the EMI primarily presents as asynchronous interference, manifesting as spokes, spirals, and speckles, negatively impacting data quality and application. Therefore, to improve the quality of radar signals received by weather radar, it is necessary to detect whether the radar signal is subject to EMI. Currently, some EMI identification algorithms are based on in-phase / quadrature (I / Q) data, while others are based on fundamental data. I / Q data consists of complex signals with a 90-degree phase shift in the radar echo. The base data consists of parameters such as reflectivity, velocity, and spectral width calculated based on I / Q data.

[0022] The main electromagnetic interference (EMI) identification algorithms based on I / Q data are SQI (Signal Quality Indicator) and radial noise algorithms. The SQI algorithm utilizes the characteristic that the SQI is less than 0.5 when EMI is present, using a threshold to identify EMI. If the SQI of a target cell is less than this threshold, EMI is considered present. However, since the SQI values ​​of EMI vary between 0 and 0.5 across different stations and ranges, selecting a single threshold for discrimination across the entire detection range can easily lead to missed EMI detections and misjudgments of useful signals. Furthermore, the SQI during severe convective weather can also be less than 0.5, and using this as a criterion could harm weather signals. The radial noise algorithm utilizes the characteristic that noise increases significantly when radial interference is present, estimating the noise in each radial direction in real time to determine the presence of EMI. However, while this method is effective for radial interference, it is ineffective for spiral or speckled interference. Therefore, the accuracy of existing EMI identification methods is relatively low.

[0023] Therefore, in view of the above problems, this application provides an electromagnetic interference identification method, apparatus, device and readable storage medium to achieve accurate identification of various electromagnetic interferences, and can avoid missed detection of electromagnetic interference and misjudgment of useful signals, thereby improving the accuracy of radar signal identification of whether it is subject to electromagnetic interference.

[0024] Please see Figure 1 The electromagnetic interference method provided in this application includes the following steps S101 to S103.

[0025] S101. Obtain basic data of the radar signal to be identified.

[0026] In this embodiment, basic data of the radar signal to be identified can be acquired first. This basic data consists of feature data that characterizes the features or attributes of the radar signal to be identified, and may include in-phase orthogonal data of the radar signal to be identified in multiple range databases. Here, a range database is a small range unit divided radially according to distance in radar echo signal processing. That is, the radar signal to be identified can be divided into multiple range databases according to distance, and in-phase orthogonal data (I / Q data) of the radar signal to be identified in each range database can be acquired. It is understood that the radar signal to be identified can be multiple radar signals; that is, multiple radar signals can be identified simultaneously.

[0027] S102. Calculate the characteristic parameters of the radar signal to be identified based on the basic data.

[0028] In this embodiment, after acquiring the basic data of the radar signal to be identified, the characteristic parameters of the radar signal to be identified can be calculated based on the basic data, so as to determine the identification result of the radar signal to be identified through the electromagnetic interference identification model.

[0029] In one possible implementation, the maximum power variation coefficient, skewness coefficient, kurtosis coefficient, and maximum sharpness of the radar signal to be identified can be calculated based on the I / Q data of the radar signal to be identified in multiple range databases.

[0030] Specifically, if the number of radar signals to be identified is M, and the total number of range databases for the radar signals to be identified is N, then the I / Q data of the i-th radar signal in the k-th range database can be expressed as: x(k) i , i=1,2…M, k=1,2…N; Where, x(k) i This represents the I / Q data of the i-th radar signal in the k-th range database.

[0031] First, calculate the power data of the radar signal to be identified. The calculation formula is as follows: X(k) i =x(k) i conj(x(k) i ); Where, X(k) i This represents the power data of the i-th radar signal in the k-th range database, and conj() is the conjugate operation.

[0032] The formula for calculating the maximum power variation coefficient Maxv is: MaxV(k) = Max(X) - Min(X); Where MaxV(k) is the maximum power change coefficient of the k-th distance reservoir; X is X(k). i , abbreviated as , represents the power data of the i-th radar signal in the k-th range database.

[0033] The formula for calculating the skewness coefficient Sk is: ; Where Sk(k) is the skewness coefficient of the k-th distance database; E is the expected value of the I / Q data in all distance databases; σ is the sample standard deviation; and μ is the mean.

[0034] The formula for calculating the kurtosis coefficient Ku is: ; Where Ku(k) is the kurtosis coefficient of the k-th distance library.

[0035] The operator Lap(n) can be defined. A typical operator for n=3 is as follows: .

[0036] The formula for calculating the maximum sharpness based on the operator Lap(n) is: Macu(k) = max(conv(X) k ,Lap)); Where Macu(k) is the maximum sharpness of the k-th distance library; X k =10 log 10 (abs(x(i) k )), i = 0, 1, ... M-1; conv represents convolution, abs represents absolute value.

[0037] S103. Based on the characteristic parameters and the electromagnetic interference identification model, determine the identification result of the radar signal to be identified.

[0038] In this embodiment, after calculating the feature parameters of the radar signal to be identified, the feature parameters can be input into the electromagnetic interference (EMI) identification model, which then outputs the identification result of the radar signal. It is understood that since the training dataset used to train the EMI identification model includes the feature parameters of the radar signal and the corresponding identification results, inputting the feature parameters of the radar signal to be identified into the trained EMI identification model will yield an accurate identification result.

[0039] Specifically, the training process of the electromagnetic interference identification model includes: acquiring a training dataset and a validation dataset. The training dataset includes the training feature parameters and the training identification results of the radar signal to be identified, and the validation dataset includes the verification feature vector and the verification identification results of the radar signal to be identified; creating a neural network model; inputting the training dataset into the neural network model for training to obtain an initial electromagnetic interference identification model; and adjusting the performance parameters of the initial electromagnetic interference identification model according to the validation dataset to obtain the electromagnetic interference identification model.

[0040] Understandably, one can first select basic data of multiple radar signals subject to electromagnetic interference and multiple radar signals not subject to electromagnetic interference from the historical database, calculate their feature parameters respectively, use the feature parameters and the corresponding recognition results as training sample data, and randomly divide the training samples into training set and validation set according to the M:1 ratio.

[0041] Then, create a neural network model, setting the preset number of hidden layers, the number of neurons in each hidden layer, the activation function and the number of output parameters for each layer, selecting the loss function and optimizer. The created neural network model looks like this. Figure 2 As shown. This neural network model has N hidden layers. The first N-1 layers are constructed using fully connected layers, and the activation function used is tanh, which is calculated as follows: .

[0042] To prevent overfitting during training, the Dropout algorithm is used to prevent the activation of some neurons. The output layer does not include an activation function. Since the distinction between electromagnetic interference and non-electromagnetic interference needs to be made, the number of output parameters is 2. If more targets need to be identified, the number of output parameters can be adjusted accordingly. Softmax is used to transform the output into probabilities for use in calculating the loss function.

[0043] The loss function can be cross-entropy, and the formula for cross-entropy is: ; Where y is the true probability distribution. To predict probability distributions, cross-entropy describes the distance between two probability distributions; the smaller the cross-entropy value, the better the prediction result.

[0044] The optimizer used is Adam (Adaptive Moment Estimation), which uses the first moment estimate (mean gradient) and second moment estimate (variance of uncentered gradient) of the loss function gradient to dynamically adjust the learning rate α of each parameter (W,b).

[0045] The backpropagation algorithm is used to update parameters W and b. The main feature of the backpropagation algorithm is that the signal is propagated forward while the error is propagated backward. By continuously adjusting the weight values, the final output of the network is made as close as possible to the expected output, so as to achieve the training objective. The parameter update formula is as follows: ; ; Where α is the learning rate, and its value ranges from (0,1); J(W,b) is the loss function; This represents the connection weight from the j-th neuron in layer L-1 to the i-th neuron in layer L; It is the bias of the i-th neuron in the L-th layer.

[0046] The initial learning efficiency is set to a constant, the initial weights are set to an all-zero matrix, the initial bias is set to an all-zero vector, and the preset number of iterations is N.

[0047] After creating the neural network model, the training dataset can be input into the model, and the neural network can be trained using the batch gradient descent training method to obtain the trained weights W and biases b, resulting in the initial electromagnetic interference model. Then, the training set and validation set data are input together into the adjusted optimal model for training, finding the optimal function that minimizes the loss function, obtaining the final model parameters, and finally obtaining the trained electromagnetic interference identification model.

[0048] In one possible implementation, feature parameters can be input into the electromagnetic interference (EMI) identification model to obtain the identification results of the radar signal to be identified in multiple range databases. It is understood that after inputting the feature parameters of the radar signal to be identified into the EMI identification model, the model can output the identification results of the radar signal to be identified in each range database, that is, it can determine whether the radar signal to be identified is subject to electromagnetic interference in each range database.

[0049] In one possible implementation, the contaminated area of ​​the radar signal to be identified can be determined based on the identification results of the radar signal in multiple range databases; the contaminated area is recorded, and an alarm is issued to the technicians. Understandably, after determining the identification results of the radar signal in multiple range databases, the contaminated area of ​​the radar signal to be identified due to electromagnetic interference can be further determined, i.e., in which range databases electromagnetic interference exists, these areas are marked and recorded, and an alarm is issued to the technicians so that the technicians can further process the portion of the radar signal to be identified that is affected by electromagnetic interference.

[0050] In summary, the embodiments of this application provide an electromagnetic interference identification method. By calculating the characteristic parameters of the radar signal to be identified based on the basic data of the radar signal to be identified and inputting them into a pre-trained electromagnetic interference identification model, it is possible to accurately identify various electromagnetic interferences and avoid missing electromagnetic interference and misjudging useful signals, thus greatly improving the accuracy of identifying whether the radar signal is subject to electromagnetic interference.

[0051] Please see Figure 3 This application also provides an electromagnetic interference identification device, including: Acquisition unit 301 is used to acquire basic data of the radar signal to be identified; The calculation unit 302 is used to calculate the feature parameters of the radar signal to be identified based on the basic data; The determining unit 303 is used to determine the identification result of the radar signal to be identified based on the feature parameters and the electromagnetic interference identification model.

[0052] Optionally, the basic data of the radar signal to be identified includes in-phase orthogonal data of the radar signal to be identified in multiple range databases.

[0053] Optionally, the calculation unit 302 is specifically used to calculate the maximum power variation coefficient, skewness coefficient, kurtosis coefficient, and maximum sharpness of the radar signal to be identified based on the in-phase orthogonal data.

[0054] Optionally, the determining unit 303 is specifically used to input the feature parameters into the electromagnetic interference identification model to obtain the identification results of the radar signal to be identified in multiple range databases.

[0055] Optionally, the determining unit 303 is further configured to determine the contaminated area of ​​the radar signal to be identified based on the identification results of the radar signal to be identified in multiple distance databases; and further configured to record the contaminated area and issue an alarm prompt to the technician.

[0056] Optionally, the training process of the electromagnetic interference identification model includes: acquiring a training dataset and a validation dataset, wherein the training dataset includes the training feature parameters and the training identification results of the radar signal to be identified, and the validation dataset includes the verification feature vector and the verification identification results of the radar signal to be identified; creating a neural network model; inputting the training dataset into the neural network model for training to obtain an initial electromagnetic interference identification model; and adjusting the performance parameters of the initial electromagnetic interference identification model according to the validation dataset to obtain the electromagnetic interference identification model.

[0057] In summary, the embodiments of this application provide an electromagnetic interference identification device. By calculating the characteristic parameters of the radar signal to be identified based on the basic data of the radar signal to be identified and inputting them into a pre-trained electromagnetic interference identification model, it can accurately identify various electromagnetic interferences and avoid missing electromagnetic interference and misjudging useful signals, thus greatly improving the accuracy of identifying whether the radar signal is subject to electromagnetic interference.

[0058] This application also provides a computer device, including: a memory, a processor, and a bus system. The memory stores a program; the bus system connects the memory and the processor to enable communication between them; the processor executes the program stored in the memory to implement the electromagnetic interference identification method described above.

[0059] This application also provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the electromagnetic interference identification method as described above.

[0060] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0061] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for identifying electromagnetic interference, characterized in that, An electromagnetic interference identification method is applied to a scenario involving co-frequency asynchronous interference in weather radar, where the interference appears as spokes, spirals, or speckles on the radar image. The method includes: Acquire basic data of the radar signal to be identified, including in-phase orthogonal data of the radar signal to be identified in multiple range databases; The maximum power variation coefficient, skewness coefficient, kurtosis coefficient, and maximum sharpness of the radar signal to be identified are calculated based on the in-phase orthogonal data of the radar signal to be identified in multiple range databases. The maximum power variation coefficient, skewness coefficient, kurtosis coefficient and maximum sharpness of the radar signal to be identified are input into a pre-trained electromagnetic interference identification model. The electromagnetic interference identification model adopts a fully connected neural network structure, and the identification results of the radar signal to be identified in multiple range databases are obtained through the output layer. The identification results are used to distinguish between useful meteorological echo signals and electromagnetic interference signals. Based on the identification results of the radar signal to be identified in multiple range databases, the contaminated area of ​​the radar signal to be identified is determined, the contaminated area is recorded, and an alarm is issued to the technicians.

2. The electromagnetic interference identification method according to claim 1, characterized in that, The calculation of the maximum power variation coefficient of the radar signal to be identified based on in-phase orthogonal data of the radar signal to be identified in multiple range databases specifically includes: The power data of the radar signal to be identified is calculated using the formula: X(k) i =x(k) i conj(x(k) i ); where x(k) i This represents the in-phase orthogonal data of the i-th radar signal in the k-th range database; conj() is the conjugate operation; X(k) i This represents the power data of the i-th radar signal in the k-th range database; The maximum power variation coefficient Maxv is calculated using the formula: MaxV(k) = Max(X) - Min(X); where X is X(k). i The abbreviation for ; MaxV(k) is the maximum power change coefficient of the k-th distance library.

3. The electromagnetic interference identification method according to claim 2, characterized in that, The calculation of the skewness coefficient and kurtosis coefficient of the radar signal to be identified based on in-phase orthogonal data in multiple range databases specifically includes: The skewness coefficient Sk is calculated using the following formula: Where E is the expected value of the in-phase orthogonal data in all distance databases; σ is the sample standard deviation, μ is the mean; and Sk(k) is the skewness coefficient of the k-th distance database. The kurtosis coefficient Ku is calculated using the following formula: Where, Ku(k) is the kurtosis coefficient of the k-th distance library.

4. The electromagnetic interference identification method according to claim 2, characterized in that, The calculation of the maximum sharpness of the radar signal to be identified based on in-phase orthogonal data in multiple range databases specifically includes: Define the operator Lap(n). The typical operator for n=3 is as follows: ; The maximum sharpness is calculated based on the operator Lap(n), and the formula is: Macu(k) = max(conv(X) k ,Lap)); where X k =10 log 10 (abs(x(k) i )), i = 0, 1, ... M-1; conv represents convolution; abs represents absolute value; Macu(k) is the maximum sharpness of the k-th distance library.

5. The electromagnetic interference identification method according to claim 1, characterized in that, The electromagnetic interference identification model adopts a fully connected neural network structure. The fully connected neural network structure has N hidden layers, and the first N-1 layers are constructed using a fully connected layer approach, with tanh as the activation function. The Dropout algorithm is used in the fully connected neural network structure to prevent some neurons from being activated during training. The output layer of the fully connected neural network structure does not include an activation function. Since it is necessary to distinguish between electromagnetic interference and non-electromagnetic interference, the number of output parameters is 2. Softmax is used to convert the output results into probabilities for the calculation of the loss function.

6. The electromagnetic interference identification method according to claim 5, characterized in that, During the training process of the electromagnetic interference identification model, the loss function is cross-entropy, the optimizer is Adam optimizer, and the backpropagation algorithm is used to update the parameter weights W and bias b. During the training process, the initial learning efficiency is set to a constant, the initial weights are set to an all-zero matrix, and the initial bias is set to an all-zero vector.

7. The electromagnetic interference identification method according to claim 5, characterized in that, The electromagnetic interference identification model adopts a fully connected neural network structure, and obtains the identification results of the radar signal to be identified in multiple range databases through the output layer, specifically including: The maximum power variation coefficient, skewness coefficient, kurtosis coefficient, and maximum sharpness of the radar signal to be identified in multiple range databases are input into the electromagnetic interference identification model. The electromagnetic interference identification model outputs the identification result of the radar signal to be identified in each range database to determine whether the radar signal to be identified is subject to electromagnetic interference in each range database.

8. An electromagnetic interference identification device, characterized in that, An electromagnetic interference identification device is used in the scenario of identifying co-frequency asynchronous interference in weather radar, wherein the co-frequency asynchronous interference appears as spokes, spirals, or speckles on the radar image. The device includes: An acquisition unit is used to acquire basic data of the radar signal to be identified, the basic data including in-phase orthogonal data of the radar signal to be identified in multiple range databases; The calculation unit is used to calculate the maximum power variation coefficient, skewness coefficient, kurtosis coefficient, and maximum sharpness of the radar signal to be identified based on the in-phase orthogonal data of the radar signal to be identified in multiple range databases. The determination unit is used to input the maximum power variation coefficient, skewness coefficient, kurtosis coefficient and maximum sharpness of the radar signal to be identified into a pre-trained electromagnetic interference identification model. The electromagnetic interference identification model adopts a fully connected neural network structure and obtains the identification results of the radar signal to be identified in multiple range databases through the output layer. The identification results are used to distinguish between useful meteorological echo signals and electromagnetic interference signals. The determining unit is further configured to determine the contaminated area of ​​the radar signal to be identified based on the identification results of the radar signal to be identified in multiple distance databases, record the contaminated area, and issue an alarm prompt to the technicians.

9. A computer device, characterized in that, include: A memory, a processor, and a bus system; wherein the memory is used to store a program; and the bus system is used to connect the memory and the processor to enable communication between the memory and the processor. The processor is used to execute the program in the memory to implement the electromagnetic interference identification method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The device stores instructions that, when executed on a computer, cause the computer to perform the electromagnetic interference identification method as described in any one of claims 1 to 7.

Images