Sea clutter data generation method and device based on intelligent learning

Abstract

The invention discloses a sea clutter data generation method and device based on intelligent learning. The method comprises the following steps: constructing a network model for generating sea clutter data, wherein the network model comprises a first network and a second network which are connected in sequence; the first network is used for generating corresponding sea clutter characteristics based on preset sea clutter parameters, and the second network is used for generating corresponding sea clutter data based on the sea clutter characteristics; training the network model based on a sea clutter data sample of a preset sea area and a sea condition state to obtain a trained network model; each data sample is provided with a sea clutter feature tag and a data tag; the sea clutter data sample comprises a sea clutter theoretical simulation data sample and a sea clutter actual measurement data sample; and inputting target sea clutter parameters into the trained network model to generate corresponding sea clutter data. According to the invention, a large amount of sea clutter data can be obtained, and the precision of the data is high.

Description

Technical Field

[0001] The present invention relates to the technical field of radar detection, and particularly relates to a method and device for generating sea clutter data based on intelligent learning. Background Art

[0002] Currently, sea clutter data mainly comes from theoretical simulation and test measurement. Among them, the theoretical simulation data of sea clutter can obtain data in a wide range of frequency bands, angles, multiple polarizations, and sea conditions, but the data accuracy is low; while the test measurement data is limited by factors such as test instruments and sea conditions, and the data volume is small, which is not enough to meet the demand for sea clutter data. Summary of the Invention

[0003] The present invention provides a method and device for generating sea clutter data based on intelligent learning, which can obtain a large amount of sea clutter data with high accuracy. The technical solutions are as follows:

[0004] On the one hand, a method for generating sea clutter data based on intelligent learning is provided. The method includes:

[0005] Construct a network model for generating sea clutter data, where the network model includes a first network and a second network connected in sequence; the first network is used to generate corresponding sea clutter features based on preset sea clutter parameters, and the second network is used to generate corresponding sea clutter data based on the sea clutter features;

[0006] Train the network model based on sea clutter data samples of a preset sea area and sea condition state to obtain a trained network model; each data sample is set with a sea clutter feature label and a data label;

[0007] Input target sea clutter parameters into the trained network model to generate corresponding sea clutter data.

[0008] On the other hand, a device for generating sea clutter data based on intelligent learning is provided. The device includes:

[0009] A construction unit for constructing a network model for generating sea clutter data, where the network model includes a first network and a second network connected in sequence; the first network is used to generate corresponding sea clutter features based on preset sea clutter parameters, and the second network is used to generate corresponding sea clutter data based on the sea clutter features;

[0010] A training unit for training the network model based on sea clutter data samples of a preset sea area and sea condition state to obtain a trained network model; each data sample is set with a sea clutter feature label and a data label;

[0011] A generating unit, configured to input target sea clutter parameters into the trained network model to generate corresponding sea clutter data.

[0012] On the other hand, a computer device is provided, which includes a memory and a processor. The memory is used to store a computer program, and the processor is configured to execute the computer program stored on the memory to implement the steps of the above-mentioned method for generating sea clutter data based on intelligent learning.

[0013] On the other hand, a computer-readable storage medium is provided. The storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned method for generating sea clutter data based on intelligent learning are implemented.

[0014] On the other hand, a computer program product is provided, including a computer program. When the computer program is executed by a processor, the steps of the above-mentioned method for generating sea clutter data based on intelligent learning are implemented.

[0015] An embodiment of the present invention provides a method for generating sea clutter data based on intelligent learning. First, a network model for generating sea clutter data is constructed based on an intelligent learning algorithm. This network model can generate sea clutter data according to sea clutter parameters. Then, the network model is trained using sea clutter theoretical simulation data samples and sea clutter measured data samples to obtain a trained network model. Therefore, the model can effectively learn all the features in the sea clutter parameters, enabling it to have a high parameter recognition ability and sea clutter data generation ability, and the generated sea clutter data has high accuracy. Finally, when the target sea clutter parameters are determined, as long as they are input into the network model, the corresponding sea clutter data can be quickly generated. Thus, it can be seen that by using the method of the present application, a large amount of sea clutter data can be quickly obtained, and the data has high accuracy. Description of the Drawings

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0017] Figure 1 is a flowchart of a method for generating sea clutter data based on intelligent learning provided by an embodiment of the present invention;

[0018] Figure 2 is a structural diagram of a device for generating sea clutter data based on intelligent learning provided by an embodiment of the present invention;

[0019] Figure 3 It is a hardware architecture diagram of a computer device provided by an embodiment of the present invention;

[0020] Figure 4 It is a schematic structural diagram of a first network provided by an embodiment of the present invention;

[0021] Figure 5 It is a schematic structural diagram of a second network provided by an embodiment of the present invention. Detailed implementation manners

[0022] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Apparently, the described embodiments are some but not all of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0023] The following describes the specific implementation manners of the above concepts.

[0024] Please refer to Figure 1 , a method for generating sea clutter data based on intelligent learning provided by an embodiment of the present invention, the method includes:

[0025] Step 100, constructing a network model for generating sea clutter data, the network model includes a first network and a second network connected in sequence; the first network is used to generate corresponding sea clutter features based on preset sea clutter parameters, and the second network is used to generate corresponding sea clutter data based on the sea clutter features;

[0026] Step 102, training the network model based on sea clutter data samples of a preset sea area and sea state, to obtain a trained network model; each data sample is set with a sea clutter feature label and a data label; the sea clutter data samples include sea clutter theoretical simulation data samples and sea clutter measured data samples;

[0027] Step 104, inputting target sea clutter parameters into the trained network model to generate corresponding sea clutter data.

[0028] In this embodiment, first, a network model for generating sea clutter data is constructed based on an intelligent learning algorithm. This network model can generate sea clutter data according to sea clutter parameters. Then, the network model is trained using sea clutter theoretical simulation data samples and sea clutter measured data samples to obtain a trained network model. Therefore, this model can effectively learn all the features in the sea clutter parameters, enabling it to have high parameter recognition ability and sea clutter data generation ability, and the generated sea clutter data has high accuracy. Finally, when the target sea clutter parameters are determined, as long as they are input into this network model, the corresponding sea clutter data can be quickly generated. It can be seen that by using the method of this application, a large amount of sea clutter data can be quickly obtained, and the accuracy of the data is high.

[0029] The execution manner of each step described below Figure 1 is shown as follows.

[0030] First, for step 100, a network model for generating sea clutter data is constructed.

[0031] In this step, the sea clutter parameters include frequency band, elevation angle, and polarization mode.

[0032] In addition, as Figure 4 shown, the first network includes a sea clutter physical simulation layer, a feature mapping layer, and a first fully connected layer connected in sequence; the output of the sea clutter physical simulation layer is used as the input of the feature mapping layer, and the output of the feature mapping layer is used as the input of the first fully connected layer; the sea clutter physical simulation layer is used to input sea clutter parameters and generate corresponding sea clutter data according to the sea clutter parameters and a preset sea surface electromagnetic scattering theoretical model; the feature mapping layer is used to realize the mapping of sea clutter data in the feature space; the first fully connected layer is used to map the feature representation to a low-dimensional sea clutter data feature space to output sea clutter data features;

[0033] As Figure 5 shown, the second network includes a second fully connected layer and a generation layer connected in sequence; the output of the feature mapping network is used as the input of the second fully connected layer, and the output of the second fully connected layer is used as the input of the generation layer; the second fully connected layer is used to perform dimensionality increase on the data representation in the sea clutter data feature space; the generation layer is used to generate and output sea clutter data according to the input features.

[0034] In addition, the feature mapping layer at least includes a convolutional layer, batch normalization, and a non-linear activation function connected in sequence. Among them, the convolutional layer includes operations of performing convolution with a convolution kernel and downsampling, where the convolution kernel takes parameters, and the downsampling multiple is determined according to needs. Batch normalization processes the data output by the convolutional layer as follows: subtracting the data mean and then dividing by the data standard deviation. The non-linear activation function uses the ReLU function or the LeakyReLU function.

[0035] The generation layer at least includes a transposed convolution layer, batch normalization, and a non-linear activation function connected in sequence. Among them, the transposed convolution layer contains operations of convolving with a convolution kernel and upsampling, where the convolution kernel parameters and the upsampling factor are determined according to needs. Batch normalization processes the data output by the transposed convolution layer as follows: subtracting the data mean and then dividing by the data standard deviation; the non-linear activation function uses the ReLU function or the LeakyReLU function.

[0036] In some embodiments, the difference in the number of layers between the feature mapping layer and the generation layer is not greater than one layer, and preferably the number of layers of the feature mapping layer is equal to the number of layers of the generation layer.

[0037] In some embodiments, the loss function of the network model is the deviation generated based on the theoretical simulation data of sea clutter The deviation generated from the measured data of sea clutter and the deviation generated from the fusion of the theoretical simulation data of sea clutter and the measured data of sea clutter in the feature space determined;

[0038] The expression of the loss function is:

[0039]

[0040] In the formula, θ E represents the sea clutter parameter corresponding to the first network, θ G represents the sea clutter parameter corresponding to the second network, x sim represents the theoretical simulation data of sea clutter, x mea represents the measured data of sea clutter, represents the first network, represents the second network.

[0041] In some embodiments, the root mean square error between the theoretical simulation data of sea clutter and the measured data of sea clutter under the same conditions is less than a preset value, such as 4 dB.

[0042] Regarding step 102, the training process of the network model is as follows:

[0043] Divide the theoretical simulation data samples of sea clutter and the measured data samples of sea clutter into a training set and a test set according to a preset ratio;

[0044] Use the training set and the test set to train the network model. For each round of training, calculate the loss function of this round until the obtained loss function is less than a preset threshold to obtain the trained network model.

[0045] In this step, the number of samples of the theoretical simulation data of sea clutter is greater than the number of samples corresponding to the measured data of sea clutter. The ratio between the two is determined according to needs, such as 30:1 to 90:1.

[0046] Finally, for step 104, as long as the target sea clutter parameters of interest, such as the target frequency band, target elevation angle, and target polarization mode, are determined, and the above parameters are input into the trained network model, the corresponding sea clutter data can be generated, greatly improving the generation speed and data accuracy.

[0047] Such as Figure 2 、 Figure 3 As shown, an embodiment of the present invention provides a sea clutter data generation device based on intelligent learning. The device embodiment can be implemented by software, or by hardware or a combination of software and hardware. From the hardware level, as Figure 2 shown, it is a hardware architecture diagram of a computing device where a sea clutter data generation device based on intelligent learning provided by an embodiment of the present invention is located. In addition to Figure 2 the processor, memory, network interface, and non-volatile memory shown, the computing device where the device is located in the embodiment usually may also include other hardware, such as a forwarding chip responsible for processing packets, etc. Taking software implementation as an example, as Figure 3 shown, as a logically meaningful device, it is formed by the CPU of its computing device reading the corresponding computer program in the non-volatile memory into the memory and running.

[0048] Please refer to Figure 3 , an embodiment of the present invention provides a sea clutter data generation device based on intelligent learning. The device includes:

[0049] A construction unit 300, configured to construct a network model for generating sea clutter data. The network model includes a first network and a second network connected in sequence. The first network is used to generate corresponding sea clutter features based on preset sea clutter parameters, and the second network is used to generate corresponding sea clutter data based on the sea clutter features;

[0050] A training unit 302, configured to train the network model based on sea clutter data samples of a preset sea area and sea condition state to obtain a trained network model; each data sample is set with a sea clutter feature label and a data label;

[0051] A generation unit 304, configured to input target sea clutter parameters into the trained network model to generate corresponding sea clutter data.

[0052] In some embodiments, the sea clutter parameters include a frequency band, an elevation angle, and a polarization mode.

[0053] In some embodiments, the first network includes a sea clutter physical simulation layer, a feature mapping layer, and a first fully-connected layer connected in sequence; the output of the sea clutter physical simulation layer serves as the input of the feature mapping layer, and the output of the feature mapping layer serves as the input of the first fully-connected layer; the sea clutter physical simulation layer is used to input sea clutter parameters and generate corresponding sea clutter data according to the sea clutter parameters and a preset sea surface electromagnetic scattering theory model; the feature mapping layer is used to implement the mapping of sea clutter data in the feature space; the first fully-connected layer is used to map the feature representation to a low-dimensional sea clutter data feature space to output sea clutter data features;

[0054] The second network includes a second fully-connected layer and a generation layer connected in sequence. The output of the feature mapping network serves as the input of the second fully-connected layer, and the output of the second fully-connected layer serves as the input of the generation layer; the second fully-connected layer is used to increase the dimension of the data representation in the sea clutter data feature space; the generation layer is used to generate and output sea clutter data according to the input features.

[0055] In some embodiments, the feature mapping layer at least includes a convolutional layer, batch normalization, and a non-linear activation function connected in sequence; the generation layer at least includes a transposed convolutional layer, batch normalization, and a non-linear activation function connected in sequence;

[0056] The difference in the number of layers between the feature mapping layer and the generation layer is not greater than one layer.

[0057] In some embodiments, the loss function of the network model is based on the deviation generated by the sea clutter theoretical simulation data The deviation generated by the measured sea clutter data And the deviation generated by the fusion of the sea clutter theoretical simulation data and the measured sea clutter data in the feature space Determined;

[0058] The expression of the loss function is:

[0059]

[0060] In the formula, θ E Represents the sea clutter parameters corresponding to the first network, θ G Represents the sea clutter parameters corresponding to the second network, x sim Represents the sea clutter theoretical simulation data, x mea Represents the measured sea clutter data, Represents the first network, Represents the second network.

[0061] In some embodiments, the root mean square error between the sea clutter theoretical simulation data and the measured sea clutter data under the same conditions is less than a preset value.

[0062] In some embodiments, the training process of the network model is as follows:

[0063] Divide the theoretical simulation data samples and measured data samples of sea clutter into a training set and a test set according to a preset ratio;

[0064] Use the training set and the test set to train the network model. For each round of training, calculate the loss function of this round until the obtained loss function is less than a preset threshold to obtain a trained network model.

[0065] It should be noted that: The sea clutter data generation device based on intelligent learning provided in the above embodiments is only illustrated by dividing the above functional modules. In practical applications, the above functions can be allocated to different functional modules according to needs, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the sea clutter data generation device based on intelligent learning provided in the above embodiments and the embodiments of the sea clutter data generation method based on intelligent learning belong to the same concept. For the specific implementation process, please refer to the method embodiments and will not be elaborated here.

[0066] The embodiments of the present application also provide a computer device. Please refer to Figure 3 . The computer device includes a processor and a memory. At least one instruction, at least one program, a code set or an instruction set is stored in the memory. At least one instruction, at least one program, a code set or an instruction set is loaded and executed by the processor to implement the sea clutter data generation method based on intelligent learning provided in the above method embodiments.

[0067] The embodiments of the present application also provide a computer-readable storage medium. At least one instruction, at least one program, a code set or an instruction set is stored on the computer-readable storage medium. At least one instruction, at least one program, a code set or an instruction set is loaded and executed by the processor to implement the sea clutter data generation method based on intelligent learning provided in the above method embodiments.

[0068] The embodiments of the present application also provide a computer program product. The computer program product includes a computer program. The processor of the computer device reads the computer program from the computer-readable storage medium, and the processor executes the computer program, so that the computer device executes the sea clutter data generation method based on intelligent learning described in any one of the above embodiments.

[0069] For the convenience of description, when describing the above system or device, it is divided into various modules or units according to functions for description. Of course, when implementing the present application, the functions of each unit can be implemented in one or more software and / or hardware.

[0070] As can be seen from the description of the above embodiments, those skilled in the art can clearly understand that this application can be implemented by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of this application, in essence, or the part that contributes to the prior art can be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in various embodiments or some parts of the embodiments of this application.

[0071] Finally, it should also be noted that in this article, relational terms such as first, second, third, and fourth are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "including a..." does not exclude the existence of additional identical elements in the process, method, article or device including the said element.

[0072] The above are only the preferred embodiments of this application. It should be pointed out that for those of ordinary skill in the art, without departing from the principle of this application, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of this application.

Claims

1. A method for generating sea clutter data based on intelligent learning, characterized in that: The method comprises: Constructing a network model for generating sea clutter data, the network model comprising a first network and a second network connected in sequence; the first network is used to generate corresponding sea clutter features based on preset sea clutter parameters, and the second network is used to generate corresponding sea clutter data based on the sea clutter features; The network model is trained based on sea clutter data samples of a preset sea area and sea state to obtain a trained network model; each data sample is provided with a sea clutter feature label and a data label; the sea clutter data samples include sea clutter theoretical simulation data samples and sea clutter measured data samples; The target sea clutter parameters are input into the trained network model to generate corresponding sea clutter data.

2. The method according to claim 1, characterized in that: The sea clutter parameters include frequency band, elevation angle and polarization mode.

3. The method according to claim 1, characterized in that The first network includes a sea clutter physical simulation layer, a feature mapping layer and a first fully connected layer connected in sequence; the output of the sea clutter physical simulation layer is used as the input of the feature mapping layer, and the output of the feature mapping layer is used as the input of the first fully connected layer; the sea clutter physical simulation layer is used to input sea clutter parameters, and generate corresponding sea clutter data according to the sea clutter parameters and a preset sea surface electromagnetic scattering theoretical model; the feature mapping layer is used to realize the mapping of sea clutter data in a feature space; the first fully connected layer is used to map the feature representation to a low-dimensional sea clutter data feature space to output sea clutter data features; The second network includes a second fully connected layer and a generating layer connected in sequence, the output of the feature mapping network serves as the input of the second fully connected layer, and the output of the second fully connected layer serves as the input of the generating layer; the second fully connected layer is used to increase the dimension of the data representation of the sea clutter data feature space; the generating layer is used to generate and output the sea clutter data according to the input features.

4. The method according to claim 3, characterized in that The feature mapping layer at least includes a convolution layer, a batch normalization, and a nonlinear activation function connected in sequence; the generation layer at least includes a deconvolution layer, a batch normalization, and a nonlinear activation function connected in sequence; The difference in the number of levels between the feature mapping layer and the generation layer is no more than one level.

5. The method according to claim 1, characterized in that The loss function of the network model is based on the deviation of the sea clutter theory simulation data. Deviations from measured sea clutter data As well as the deviation caused by the fusion of sea clutter theoretical simulation data and sea clutter measured data in feature space Determined; The expression of the loss function is: In the formula, θ E represents the sea clutter parameter corresponding to the first network, θ G represents the sea clutter parameter corresponding to the second network, x sim represents the theoretical simulation data of sea clutter, x mea represents the measured data of sea clutter, represents the first network, Indicates the second network.

6. The method according to claim 5, characterized in that The root mean square error between the theoretical simulation data of sea clutter and the measured data of sea clutter under the same conditions is less than a preset value.

7. The method according to claim 1, characterized in that The training process of the network model is as follows: Dividing the sea clutter theoretical simulation data samples and the sea clutter measured data samples into a training set and a test set according to a preset ratio; The network model is trained using the training set and the test set. For each round of training, the loss function of that round is calculated until the obtained loss function is less than a preset threshold, thereby obtaining a trained network model.

8. A sea clutter data generation device based on intelligent learning, characterized in that: The device comprises: A construction unit, configured to construct a network model for generating sea clutter data, the network model comprising a first network and a second network connected in sequence; the first network is configured to generate corresponding sea clutter features based on preset sea clutter parameters, and the second network is configured to generate corresponding sea clutter data based on the sea clutter features; A training unit, used to train the network model based on sea clutter data samples of a preset sea area and sea state to obtain a trained network model; each data sample is provided with a sea clutter feature label and a data label; The generating unit is used to input the target sea clutter parameters into the trained network model to generate corresponding sea clutter data.

9. A computer device, characterized in that: The computer device includes a memory and a processor, the memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory to implement the steps of any one of the methods described in claims 1-7.

10. A computer-readable storage medium, characterized in that: The storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method described in any one of claims 1 to 7 are implemented.

Images