Single-frequency feature extraction method based on full-connection network and enhanced discrete feature loss function
By adopting a single-frequency feature extraction method based on a fully connected network and enhanced discrete feature loss function in water acoustic target recognition, the problems of high false alarm rate, poor stability and large calculation overhead of single-frequency feature extraction in the prior art are solved, and efficient and accurate single-frequency feature extraction is achieved.
Patent Information
- Application Number
- CN202510196122.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-02-21
- Publication Date
- 2025-05-27
AI Technical Summary
When extracting the single-frequency characteristics of water acoustic targets, the false alarm rate is high, the stability is poor, and the computational overhead of complex neural networks is large, making it difficult to meet the real-time processing requirements.
A single-frequency feature extraction method based on a fully connected network and enhanced discrete feature loss function is adopted. By constructing a multi-layer fully connected layer network model and designing an enhanced discrete feature loss function, a feature extraction model that is sensitive to single-frequency features is trained.
Single-frequency feature extraction with less computational overhead is realized, and single-frequency features in target radiation noise can be extracted more accurately, providing a basis for target recognition.
Smart Images
Figure CN120045920A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical fields of underwater acoustic target parameter estimation and artificial intelligence technology, and mainly relates to a single-frequency feature extraction method based on a fully connected network and an enhanced discrete feature loss function. Background Art
[0002] The single-frequency feature of the underwater acoustic target radiated noise power spectrum / LOFAR spectrum is an important feature in underwater acoustic target recognition. The single-frequency feature is often related to the operation of equipment such as the main engine and auxiliary engine of the underwater acoustic target, and belongs to the inherent feature information of the underwater acoustic target, which can provide important information for target recognition.
[0003] However, for the current method of extracting the target single-frequency feature by a simple threshold method, the target radiated noise is affected by the transmission of the ocean acoustic channel and the interference of other targets, and the extracted single-frequency feature has a high false alarm rate and poor stability. At the same time, for the single-frequency feature extraction method using a complex neural network, the computational cost is large and it is difficult to meet the requirements of real-time processing. Summary of the Invention
[0004] The purpose of the present invention is to overcome the deficiencies existing in the prior art and provide a single-frequency feature extraction method based on a fully connected network and an enhanced discrete feature loss function.
[0005] The purpose of the present invention is achieved by the following technical solutions. A single-frequency feature extraction method based on a fully connected network and an enhanced discrete feature loss function includes the following steps:
[0006] S1. Construct a network model for single-frequency feature extraction based on a fully connected network;
[0007] (1.1) Construct a power spectrum / LOFAR spectrum input layer
[0008] Construct an input layer adapted to the input power spectrum / LOFAR spectrum. The number of nodes in the input layer is greater than or equal to the number of frequency points of the spectrum. When the number of nodes in the input layer is equal to the number of frequency points of the spectrum, the node values are equal to the spectrum frequency point values in sequence and are input into the network. When the number of nodes in the input layer is greater than the number of frequency points of the spectrum, the spectrum frequency points are expanded to be equal to the number of nodes in the input layer through interpolation and then input into the network;
[0009] (1.2) Construct a multi-layer single-frequency feature extraction structure
[0010] The multi-layer single-frequency feature extraction structure is composed of multiple fully connected layers. All nodes between two adjacent fully connected layers are associated, and each node is passed to the next layer after passing through the ReLU activation function;
[0011] The number of nodes in the fully connected layer is related to the number of nodes N in the input layer 输入 The number of nodes in the first layer is N 输入 / 2, and the number of nodes in the second layer is N输入 / 2 2 , and so on until the number of nodes in layer M is N M less than 100; subsequently, in layer M+1, the number of nodes is 2*N M , and so on until layer 2M-1, the number of nodes is N 输入 / 2;
[0012] (1.3) Construct an output layer adapted to the single-frequency feature label
[0013] The number of nodes in the output layer is equal to the number of input nodes N 输入 , ensuring that the number of nodes in the output result has a corresponding relationship with the number of frequency points in the input power spectrum / LOFAR spectrum, so as to determine the frequency characteristics of the single-frequency feature;
[0014] The output layer uses the Sigmod activation function, and the calculation method is as shown in the following formula:
[0015]
[0016] making the output result distributed between (0,1), which is adapted to the discrete feature label;
[0017] S2. Use the simulation data for training, and train a feature extraction model sensitive to the single-frequency feature by designing a loss function that enhances the discrete feature;
[0018] S3. Apply the network model to process the target radiated noise and extract the target single-frequency feature.
[0019] The steps for generating the target single-frequency feature label are as follows: First, separate the continuous spectrum feature and the single-frequency feature in the target power spectrum / LOFAR spectrum to obtain the single-frequency feature with the relative energy magnitude relationship. Subsequently, according to the frequency band and the number of label points, determine the corresponding position of the single-frequency feature, and according to the relative energy magnitude relationship, determine the label value of the corresponding position.
[0020] The calculation method of the loss function for enhancing the discrete feature is as follows: First, calculate the deviation between the label and the prediction result, and then use the label to enhance the deviation corresponding to the discrete part. Specifically, multiply the deviation corresponding to the non-0 position of the label by the discrete feature weight coefficient to increase its proportion in the loss function calculation; the specific calculation method is as shown below:
[0021]
[0022] where y is the label vector, f(x) is the predicted single-frequency feature, N is the vector length, and b∈[0,1) is the discrete feature weight coefficient; when the value of b is higher, the weight of the discrete feature is higher, and the loss value is more sensitive to the discrete feature.
[0023] Process the target radiated noise using a network model to extract the target single - frequency feature. The basic process is as follows:
[0024] (3.1) Generate a power spectrum / LOFAR spectrum of the underwater acoustic target at a certain resolution, determine the number of frequency bands N, and perform global normalization;
[0025] (3.2) Input the standard power spectrum / LOFAR spectrum obtained in (3.1) into the fully - connected network for single - frequency feature extraction to obtain the corresponding single - frequency feature;
[0026] (3.3) Process the newly acquired data according to (3.1) - (3.2) to obtain the continuous result of target single - frequency feature extraction.
[0027] The beneficial effects of the present invention are as follows: The present invention provides a method for extracting single - frequency features based on a fully - connected network and an enhanced discrete feature loss function. By designing a fully - connected neural network adapted to the data and training it using the enhanced discrete feature loss function, a neural network for single - frequency feature extraction with relatively small computational overhead is obtained, which can accurately extract the single - frequency features contained in the target radiated noise and provide a basis for target recognition. The present invention is used for extracting single - frequency features from power spectra and LOFAR spectra and achieves good results. BRIEF DESCRIPTION OF THE DRAWINGS
[0028] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.
[0029] Figure 1 It shows a schematic diagram of the typical power spectrum / LOFAR spectrum of the radiated noise of an underwater acoustic target.
[0030] Figure 2 It shows a schematic diagram of the process for generating the target single - frequency feature label adopted by the present invention.‘
[0031] Figure 3 It shows a network model diagram for extracting the single - frequency feature of an underwater acoustic target.
[0032] Figure 4 It shows a schematic diagram of the calculation process of the enhanced discrete feature loss function of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS
[0033] 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. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. 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.
[0034] As Figure 1 shown in the schematic diagram of the typical power spectrum / LOFAR spectrum of underwater acoustic target radiation noise, it is mainly composed of continuous spectrum features and single-frequency features. The single-frequency features have concentrated energy and good recognition characteristics, reflecting the working state and inherent characteristics of the underwater acoustic target.
[0035] As Figure 2 shown in the target single-frequency feature label generation process adopted by the present invention, first, the continuous spectrum features and single-frequency features in the target power spectrum / LOFAR spectrum are separated to obtain single-frequency features with relative energy magnitude relationships. Subsequently, according to the frequency band and the number of label points, the corresponding positions of the single-frequency features are determined, and the label values at the corresponding positions are determined according to the relative energy magnitude relationships.
[0036] The present invention provides a single-frequency feature extraction method based on a fully connected network and an enhanced discrete feature loss function, and the specific steps are as follows:
[0037] (1) Based on a fully connected network, a network model for single-frequency feature extraction is constructed, and the basic process is as follows:
[0038] (1.1) Construct a power spectrum / LOFAR spectrum map input layer:
[0039] An input layer adapted to the input power spectrum / LOFAR spectrum map is constructed. To ensure that as much information as possible of the input power spectrum / LOFAR spectrum is preserved, the number of nodes in the input layer is greater than or equal to the number of frequency points of the spectrum map. When the number of nodes in the input layer is equal to the number of frequency points of the spectrum map, the node values are equal to the spectrum map frequency point values in sequence and are input into the network. When the number of nodes in the input layer is greater than the number of frequency points of the spectrum map, through interpolation, the number of spectrum map frequency points is expanded to be equal to the number of nodes in the input layer and then input into the network.
[0040] (1.2) Construct a multi-layer single-frequency feature extraction structure
[0041] The multi-layer single-frequency feature extraction structure is composed of multiple fully connected layers. All nodes between two adjacent fully connected layers are associated, and each node is passed to the next layer after passing through the ReLU activation function.
[0042] The number of nodes in the fully connected layer is related to the number of nodes N_input in the input layer. The number of nodes in the first layer is N_input / 2, the number of nodes in the second layer is N_input / 2^2, and so on until the number of nodes N_M in the Mth layer is less than 100. Subsequently, in the M+1th layer, the number of nodes is 2*N_M, and so on until the number of nodes in the 2M-1th layer is N_input / 2.
[0043] (1.3) Construct an output layer adapted to the single-frequency feature label
[0044] The number of nodes in the output layer is equal to the number of input nodes N_input. This ensures that the number of nodes in the output result has a corresponding relationship with the number of frequency points in the input power spectrum diagram
[0045] / LOFAR spectrum diagram frequency points, thereby determining the frequency characteristics of the single-frequency feature.
[0046] The output layer uses the Sigmod activation function, and the calculation method is shown as follows:
[0047]
[0048] This makes the output result distributed between (0,1), which is adapted to the discrete feature label;
[0049] Such as Figure 3 Shown is the network model diagram for extracting the single-frequency feature of the underwater acoustic target, which is mainly composed of an input layer, multi-layer feature extraction, and an output layer. Among them, the number of nodes in the input layer is greater than or equal to the number of frequency points of the spectrum diagram, saving all input information; the multi-layer single-frequency feature extraction structure is composed of multiple fully connected layers. All nodes between two adjacent fully connected layers are related, and each node is passed to the next layer after passing through the ReLU activation function. The number of nodes in the output layer is equal to the number of input nodes N. This ensures that the number of nodes in the output result has a corresponding relationship with the number of frequency points in the input power spectrum diagram / LOFAR spectrum diagram, thereby determining the frequency characteristics of the single-frequency feature.
[0050] (2) Design an enhanced discrete feature loss function for training the single-frequency feature extraction network model
[0051] Such as Figure 4 Shown is the calculation process of the enhanced discrete feature loss function designed by the present invention. First, calculate the deviation between the label and the prediction result, and then use the label to enhance the deviation corresponding to the discrete part. Specifically, multiply the deviation corresponding to the non-zero position of the label by the discrete feature weight coefficient, thereby increasing the proportion in the loss function calculation and improving the sensitivity of the loss function to the discrete feature.
[0052] The single-frequency feature extraction result is a 1D vector, and its length is proportional to the length and resolution of the corresponding frequency band of the power spectrum. Since the number of single-frequency features in the target radiation noise generally ranges from several to dozens, most of the elements in the label vector corresponding to the single-frequency features are 0 (no single-frequency features), and a small number of elements are greater than 0 (there are single-frequency features). In order to increase the proportion of elements with single-frequency features in the loss value, enhance the sensitivity of the loss value to single-frequency features, and improve the training efficiency of the neural network model, a loss function for enhancing discrete features is proposed, and the calculation method is as follows:
[0053]
[0054] Where y is the label vector, f(x) is the predicted single-frequency feature, N is the vector length, and b∈[0,1) is the discrete feature weight coefficient. The higher the value of b, the higher the weight of the discrete feature, and the more sensitive the loss value is to the discrete feature.
[0055] (3) Use the network model to process the target radiation noise and extract the target single-frequency features. The basic process is as follows.
[0056] (3.1) Generate the power spectrum / LOFAR spectrum of the underwater acoustic target at a certain resolution, determine the number of frequency bands N, and perform global normalization.
[0057] (3.2) Input the standard power spectrum / LOFAR spectrum obtained in (3.1) into the fully connected network for single-frequency feature extraction to obtain the corresponding single-frequency features.
[0058] (3.3) Process the newly obtained data according to (3.1) to (3.2) to obtain the continuous result of target single-frequency feature extraction.
[0059] As described above, it is only the specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the scope disclosed by the present invention should be covered by the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.
Claims
1. A single frequency feature extraction method based on a fully connected network and an enhanced discrete feature loss function, characterized in that: The following steps are involved: S1. Build a network model for single-frequency feature extraction based on a fully connected network; (1.1) Constructing the power spectrum / LOFAR spectrogram input layer Construct an input layer that is compatible with the input power spectrum / LOFAR spectrogram, and the number of input layer nodes is greater than or equal to the number of frequency points of the spectrogram; when the number of input layer nodes is equal to the number of frequency points of the spectrogram, the node values are sequentially equal to the frequency points of the spectrogram and input into the network; when the number of input layer nodes is greater than the number of frequency points of the spectrogram, the frequency points of the spectrogram are expanded to be equal to the number of input layer nodes through interpolation and input into the network; (1.2) Constructing a multi-layer single-frequency feature extraction structure The multi-layer single-frequency feature extraction structure consists of multiple layers of fully connected layers. All nodes between two adjacent fully connected layers are associated, and each node is passed to the next layer after the ReLU activation function. The number of nodes in the fully connected layer and the number of nodes in the input layer N 输入 The number of nodes in the first layer is N. 输入 / 2, the number of nodes in the second layer is N 输入 / 2 2 , and so on until the number of nodes in the M layer is N M Less than 100; then M+1 layer, the number of nodes is 2*N M , and so on until the 2M-1 layer, the number of nodes is N 输入 / 2; (1.3) Construct an output layer that is compatible with the single-frequency feature label The number of output layer nodes and the number of input nodes N 输入 Equal to ensure that the number of nodes in the output result corresponds to the number of frequency points in the input power spectrum / LOFAR spectrum, thereby determining the frequency characteristics of the single-frequency feature; The output layer uses the Sigmod activation function, and the calculation method is shown in the following formula: Make the output result distributed between (0,1) and adapt to the discrete feature label; S2. Use simulation data for training, and design a loss function that enhances discrete features to obtain a feature extraction model that is sensitive to single-frequency features; S3. Use the network model to process the target radiation noise and extract the target single-frequency characteristics.
2. The single frequency feature extraction method based on a fully connected network and an enhanced discrete feature loss function according to claim 1, characterized in that: The target single-frequency feature label generation step is: first, the continuous spectrum feature and the single-frequency feature in the target power spectrum / LOFAR spectrum are separated to obtain the single-frequency feature by the relative energy size relationship, and then the corresponding position of the single-frequency feature is determined according to the frequency band and the number of label points, and the corresponding position label value is determined according to the relative energy size relationship.
3. The single frequency feature extraction method based on a fully connected network and an enhanced discrete feature loss function according to claim 2, characterized in that: The loss function calculation method for enhancing discrete features is: first calculate the deviation between the label and the prediction result, and then use the label to enhance the deviation corresponding to the discrete part. Specifically, the deviation corresponding to the non-zero position of the label is multiplied by the discrete feature weight coefficient, thereby increasing the proportion in the loss function calculation; the specific calculation method is as follows: Where y is the label vector, f(x) is the predicted single-frequency feature, N is the vector length, and b∈[0,1) is the discrete feature weight coefficient; when the b value is higher, the weight of the discrete feature is higher, and the loss value is more sensitive to the discrete feature.
4. The single frequency feature extraction method based on a fully connected network and an enhanced discrete feature loss function according to claim 3, characterized in that: The network model is used to process the target radiation noise and extract the target single-frequency characteristics. The basic process is as follows: (3.1) Generate a power spectrum / LOFAR spectrum of the underwater acoustic target at a certain resolution, determine the number of frequency bands N, and perform global normalization; (3.2) Input the standard power spectrum / LOFAR spectrum obtained in (3.1) into the single-frequency feature extraction fully connected network to obtain the corresponding single-frequency feature; (3.3) Process the newly acquired data according to (3.1) to (3.2) to obtain continuous results of target single-frequency feature extraction.