Underwater passive positioning method combining multi-scale feature extraction and transfer learning

Abstract

The invention discloses an underwater passive positioning method combining multi-scale feature extraction and transfer learning. The underwater passive positioning method comprises the following steps: acquiring sound wave data of a sound source to be positioned in a simulation environment; performing feature extraction on the to-be-positioned sound source sound wave data to obtain sound source sound wave features; the sound source sound wave features are input into a positioning prediction model, the sound source distance and the sound source depth are predicted, the positioning prediction model is obtained through training of a training set, and the training set is waveform data of different frequency sound pressures; and domain migration is carried out on the sound source sound wave positioning problem in the simulation environment to a real sea test environment, and underwater sound source positioning is carried out. According to the invention, the acoustic diversity and dynamic change characteristics of the marine environment can be comprehensively captured, and the adaptability of the model to the real environment is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of underwater passive positioning, and in particular relates to an underwater passive positioning method combining multi-scale feature extraction and transfer learning. Background Art

[0002] In recent years, with the continuous progress of underwater acoustic signal processing technology, underwater passive positioning, as an important branch of the underwater acoustic field, has received widespread attention. This technology determines the position of underwater targets by analyzing the sound pressure signals received by the hydrophone array. However, due to the complexity of the underwater propagation environment, such as multipath effects, time delays, and interference from environmental noise, the distance and depth estimation of the sound source is always a very challenging task. Despite this, underwater passive positioning technology has shown important value in the fields of ocean detection, seabed resource exploration, marine environment monitoring, and military applications, and has become a research hotspot in the field of underwater acoustic signal processing.

[0003] At present, the research focuses on the following aspects: improving the accuracy of sound source localization, enhancing the model's adaptability to complex environments, and developing advanced algorithms that can cope with changing ocean conditions. These research directions not only aim to overcome the technical challenges brought by the underwater acoustic environment, but also play an important role in further improving the reliability and practicality of underwater passive positioning technology.

[0004] In the existing technical system, underwater passive positioning methods are becoming increasingly diverse, mainly including traditional matching field processing, machine learning positioning, and emerging positioning algorithms based on deep learning. Each method has a specific theoretical basis, applicable scenarios and technical characteristics, and has achieved remarkable results in different underwater environments.

[0005] The matched field processing (MFP) method determines the location of the sound source by comparing the correlation between the copy field data calculated by the sound propagation model and the measured acoustic data. However, this method is highly dependent on the accuracy of the sound propagation model, and its performance may be significantly reduced when there is a mismatch between the copy field and the actual sound field environment. In addition, the complexity and dynamic changes of the underwater environment (such as fluctuations in the sound velocity profile, irregularities in the seabed topography, and background noise interference) further aggravate the model mismatch problem, limiting the applicability of the MFP method in a changing environment.

[0006] With the rapid development of machine learning and deep learning technologies, they have shown great potential in the field of underwater passive positioning. In particular, deep learning-based algorithms have made significant progress in dealing with complex acoustic environments and improving positioning accuracy. However, the application of deep learning methods still faces the major challenge of lack of labeled data. Current research mainly relies on the prior knowledge of acoustic experts to generate samples, but data scarcity greatly limits the promotion and application of these methods.

[0007] Therefore, there is an urgent need for an underwater passive positioning method that combines multi-scale feature extraction and transfer learning. Summary of the invention

[0008] In order to solve the above technical problems, the present invention proposes an underwater passive positioning method that combines multi-scale feature extraction and transfer learning, which can improve the model's ability to extract acoustic features and adapt to the real environment, thereby expanding the technology to have an important impact in practical applications.

[0009] The present invention provides an underwater passive positioning method combining multi-scale feature extraction and transfer learning, comprising:

[0010] Acquire the sound wave data of the sound source to be located in the simulation environment;

[0011] Extracting features from the sound wave data of the sound source to be located to obtain sound wave features of the sound source;

[0012] Inputting the sound source sound wave characteristics into a positioning prediction model to predict the sound source distance and sound source depth, wherein the positioning prediction model is obtained by training a training set, and the training set is waveform data of sound pressure at different frequencies;

[0013] The sound source acoustic wave localization problem in the simulation environment is domain-migrated to a real sea trial environment to perform underwater sound source localization.

[0014] Optionally, obtaining the training set includes:

[0015] Simulate the ocean environment and the waveform data of sound pressure at different frequencies within the preset bandwidth;

[0016] Superimposing the waveform data to obtain simulation waveform data;

[0017] The simulation waveform data is preprocessed to obtain the training set.

[0018] Optionally, performing feature extraction on the sound wave data of the sound source to be located to obtain the sound wave features of the sound source includes:

[0019] Performing Fourier transform on the simulation waveform data to obtain complex sound pressure data;

[0020] Normalizing the complex sound pressure data using an L2 normal form to obtain the characteristic signal;

[0021] A covariance operation is performed on the characteristic signal samples to obtain the sound wave characteristics of the sound source.

[0022] Optionally, the method of performing feature extraction on the sound wave data of the sound source to be located to obtain the sound wave features of the sound source is:

[0023] p(f)=[p1(f),p2(f),...,p L (f)] T

[0024] Where p(f) is the array signal with frequency f, p L (f) is the complex sound pressure signal of channel L at frequency f, where L is the number of vertical linear arrays of hydrophones;

[0025] The method of normalizing the complex sound pressure data using the L2 norm is:

[0026]

[0027] in, is the normalized array signal;

[0028] Then the vector obtained from the above formula is normalized for covariance processing, preliminary feature extraction is performed on the data, and the covariance matrix C(f) is constructed:

[0029]

[0030] Where C(f) is the L×L complex covariance matrix (SCM), for The transposed conjugate matrix of .

[0031] Optionally, the positioning prediction model includes: an input module, a parameter sharing module, a feature processing module and an output module;

[0032] The input module is used to input the feature data to be located;

[0033] The parameter sharing module is used for feature sharing between different tasks;

[0034] The feature processing module is used to perform distance feature processing and depth feature processing;

[0035] The output module is used to output the sound source distance and the sound source depth.

[0036] Optionally, the parameter sharing module includes: a first unit, a second unit, a convolutional attention mechanism unit and a third unit;

[0037] The first unit is used to realize multi-scale feature extraction of the model;

[0038] The second unit is used to reduce the spatial dimension of the feature map while retaining key features of multiple scales;

[0039] The convolutional attention mechanism unit is used to distinguish key features in the channel dimension and locate important areas in the spatial dimension;

[0040] The third unit is used to perform multi-scale feature extraction again based on the features processed by the convolutional attention mechanism unit to achieve feature sharing between different tasks.

[0041] Optionally, the feature processing module includes: a distance feature processing submodule and a depth feature submodule in parallel;

[0042] The distance feature processing submodule is used to process distance features;

[0043] The depth feature processing submodule is used to process depth features.

[0044] Optionally, the distance feature processing submodule includes: a fourth unit, a fifth unit and a global average pooling unit;

[0045] The fourth unit is used for multi-scale feature extraction and dimensionality reduction in the deep network;

[0046] The fifth unit is used to capture more complex local and global features and enhance the ability to combine features between branches;

[0047] The global average pooling unit is used to perform adaptive two-dimensional average pooling on the processed feature map, and finally adjust the output feature map to a specified target size.

[0048] Optionally, the deep feature processing submodule includes a fourth unit, a fifth unit, a convolutional attention mechanism unit and a global average pooling unit;

[0049] The fourth unit is used for multi-scale feature extraction and dimensionality reduction in the deep network;

[0050] The fifth unit is used to capture more complex local and global features and enhance the ability to combine features between branches;

[0051] The convolutional attention mechanism unit is used to distinguish key features in the channel dimension and locate important areas in the spatial dimension;

[0052] The global average pooling unit is used to perform adaptive two-dimensional average pooling on the processed feature map, and finally adjust the output feature map to a specified target size.

[0053] Compared with the prior art, the present invention has the following advantages and technical effects:

[0054] The present invention proposes a novel architecture for underwater target positioning in shallow sea time-varying environment. By integrating multiple advanced technologies, this architecture effectively addresses the dual challenges of lack of experimental data and environmental complexity.

[0055] The present invention is based on a multi-scale feature extraction strategy. By analyzing and extracting acoustic features at different scales, it can comprehensively capture the acoustic diversity and dynamic change characteristics of the marine environment, thereby improving the model's adaptability to the real environment. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0057] Figure 1 is a flow chart of an underwater passive positioning method combining multi-scale feature extraction and transfer learning according to an embodiment of the present invention;

[0058] Figure 2 This is a simulation experiment scene diagram of an embodiment of the present invention;

[0059] Figure 3 is a marine environment map of an embodiment of the present invention;

[0060] Figure 4 It is a diagram of the prediction model structure of an embodiment of the present invention. DETAILED DESCRIPTION

[0061] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0062] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0063] In order to deal with the problem of data scarcity, generating simulation data has become one of the solutions. By combining marine environmental parameters with acoustic tools, a large amount of simulation data can be generated for model training, and domain migration can be achieved with the help of transfer learning technology, thereby gradually reducing dependence on real data. On this basis, based on the strategy of multi-scale feature extraction, by analyzing and extracting acoustic features at different scales, the acoustic diversity of the marine environment and its dynamic change characteristics can be fully captured. Compared with traditional feature extraction methods, multi-scale feature extraction can not only provide more comprehensive acoustic information, but also improve the robustness and accuracy of the model in complex underwater acoustic environments. The application of multi-scale feature extraction methods has become an important technical means to solve the challenges of complex underwater acoustic environments.

[0064] This embodiment proposes an underwater passive positioning method combining multi-scale feature extraction and transfer learning, such as Figure 1As shown, the specific steps include:

[0065] Acquire the sound wave data to be located;

[0066] Input the sound wave data to be located into the positioning prediction model to predict the distance and depth of the sound source, wherein the positioning prediction model is obtained by training a training set, and the training set is waveform data of sound pressure at different frequencies;

[0067] Domain migration is performed on the sound source distance and depth for underwater positioning.

[0068] Specifically, for the problem of underwater passive positioning, this embodiment proposes a novel underwater positioning architecture, which uses multi-task learning technology to collaboratively predict the distance and depth of underwater sound sources. The framework diagram of the method is as follows: Figure 1 As shown in the figure, it consists of four key steps: data simulation, data preprocessing, model training, and transfer learning.

[0069] Furthermore, obtaining a training set includes:

[0070] Simulate the ocean environment and the waveform data of sound pressure at different frequencies within the preset bandwidth;

[0071] The waveform data are superimposed to obtain simulation waveform data;

[0072] Preprocess the simulation waveform data to obtain the training set.

[0073] Specifically, we first simulate the experimental scene using known environmental parameters, such as acoustic profile, sediment layer parameters, and seabed parameters. Figure 2 As shown in the figure, it shows the parameters of the simulated sound source and the arrangement of the receiving hydrophones. The ocean environment diagram is as follows Figure 3 As shown, the env file and flp file are written through the above-mentioned ocean environment, and then the Kraken toolbox is used to perform acoustic modeling of the ocean environment. The simulated waveform data of sound pressure at different frequencies within a certain bandwidth are superimposed to obtain the hydrophone receiving waveform in the time domain, and a large amount of simulation data is generated for model training.

[0074] Furthermore, the simulation waveform data is preprocessed to obtain a training set including:

[0075] Perform Fourier transform on the simulation waveform data to obtain complex sound pressure data;

[0076] The complex sound pressure data is normalized using the L2 norm to obtain the training set.

[0077] Specifically, since the target signals received by each hydrophone in the hydrophone array are correlated, while the received noise is basically uncorrelated, we can use this characteristic to perform preliminary feature extraction. First, calculate the sound pressure data received by the hydrophone, and use the normalized sample covariance matrix to consider the correlation of the signal and the uncorrelation of the noise. This method can be used as a preliminary feature extraction of the original data, effectively highlighting the target signal characteristics while suppressing the influence of uncorrelated noise. The specific steps are as follows:

[0078] The sound pressure data is expressed as p = [p1, p2, ..., p f ,…,p F ], where p F It is obtained by Fourier transforming the original sound pressure data of the sensor at frequency F. The sound pressure model p at frequency f f for:

[0079] p f =S(f)g(f,r)+ε

[0080] Among them, S(f) is the sound source term, g(f,r) is the Green's function, and ε is the noise.

[0081] Assuming that the number of vertical linear arrays of hydrophones is L, the complex sound pressure data is obtained by Fourier transforming the time domain signal p(t) received by the hydrophone, which is expressed as:

[0082] p(f)=[p1(f),p2(f),...,p L (f)] T

[0083] Considering that the intensity of the same sound source will change at different distances and depths, in order to improve the accuracy and convergence speed of the model and more accurately locate the sound source, this embodiment normalizes the data in the preprocessing stage to reduce the possible impact of random noise on the amplitude. Here, the L2 norm is used to normalize the complex sound pressure data.

[0084] The L2 normal form equation is as follows:

[0085]

[0086] Then the vector obtained by the above formula is normalized for covariance processing, and preliminary feature extraction is performed on the data to obtain the covariance matrix C(f):

[0087]

[0088] Where C(f) is the L×L complex covariance matrix (SCM), and then the real and imaginary parts of the matrix are separated to obtain L×L×2 single-frequency features. For broadband data, SCMs of different frequencies will be stacked in the third dimension of the matrix, so the network input feature data shape is L×L×2F, where F is the frequency number.

[0089] Furthermore, the positioning prediction model includes: an input module, a parameter sharing module, a feature processing module and an output module;

[0090] An input module, used for inputting data to be located;

[0091] Parameter sharing module, used for feature sharing;

[0092] A feature processing module, used for distance feature processing and depth feature processing;

[0093] Output module, used to output the sound source distance and sound source depth.

[0094] Specifically, in order to effectively handle the complex situation where the target sound source is constantly moving, this embodiment regards the sound source localization problem as a regression task. The localization algorithm model proposed in this embodiment adopts a multi-task architecture, and its model structure diagram is as follows: Figure 4 As shown in the figure, the entire network consists of four parts. The input features first enter the model from the input stream, then pass through the parameter sharing layer for feature sharing, and then predict the sound source distance and depth respectively through the distance feature branch and the depth feature branch.

[0095] Further, the parameter sharing module includes: InceptionA unit (first unit), InceptionD unit (second unit), convolutional attention mechanism unit (CBAM) and InceptionB unit (third unit);

[0096] InceptionA unit adopts a multi-branch design. Branch 1 uses 1x1 convolution to achieve dimensionality reduction and extract local features. Then branches 2 and 3 use 3x3 and 5x5 convolution operations respectively, so that the model can capture features of different receptive fields. Finally, the maximum pooling and 1x1 convolution are combined to extract pooling features, realizing efficient multi-scale feature extraction of the model.

[0097] InceptionD unit is mainly used to reduce the spatial dimension of feature maps while retaining key information for efficient downsampling. This unit uses 3x3 convolution in branches for complex downsampling to reduce the size of feature maps while retaining features of multiple scales;

[0098] Convolutional attention mechanism unit, which contains spatial attention mechanism and channel attention mechanism. This design enables the model to adaptively learn and adjust the importance of features, distinguishing key features in the channel dimension and locating important areas in the spatial dimension;

[0099] The InceptionB unit achieves efficient multi-scale feature extraction and reduces computational complexity through a multi-branch design combined with vertical and horizontal decomposition convolution operations. The difference from the InceptionA unit is that branches 2 and 3 use 1x7 and 7x1 convolution decomposition to extract long strip features, thereby increasing the receptive field while reducing computational complexity.

[0100] Further, the feature processing module includes: a parallel distance feature processing submodule and a depth feature submodule;

[0101] A distance feature processing submodule, used for processing distance features;

[0102] The deep feature processing submodule is used to process deep features.

[0103] Further, the distance feature processing submodule includes: InceptionE unit (fourth unit), InceptionC (fifth unit) unit and global average pooling unit (AdaptiveAvgPool2d);

[0104] The InceptionE unit adopts a multi-branch design, in which branch 1 uses 1x1 convolution to reduce dimension followed by 3x3 convolution to achieve downsampling, branch 2 further extracts features through vertical and horizontal decomposition convolution operations combined with 3x3 convolution, while reducing computational complexity, and branch 3 uses maximum pooling to achieve feature downsampling and retain key information. This unit performs multi-scale feature extraction and dimensionality reduction in deep networks;

[0105] The InceptionC unit adopts a multi-branch design, in which branch 1 uses 1x1 convolution to extract local features, branch 2 uses 1x3 and 3x1 convolutions to achieve efficient feature extraction and diversified representation, and branch 3 uses 3x3 convolutions followed by 1x3 and 3x1 decomposed convolutions to increase the receptive field. Through the horizontal and vertical convolution decomposition operations between branches, it captures more complex local and global features and enhances the ability to combine features between branches.

[0106] The global average pooling unit is used to perform adaptive two-dimensional average pooling on the input feature map, and finally adjust the output feature map to the specified target size.

[0107] Furthermore, the deep feature processing submodule includes: InceptionE unit, InceptionC unit, convolutional attention mechanism unit (CBAM) and global average pooling unit (AdaptiveAvgPool2d);

[0108] The InceptionE unit adopts a multi-branch design, in which branch 1 uses 1x1 convolution to reduce dimension followed by 3x3 convolution to achieve downsampling, branch 2 further extracts features through vertical and horizontal decomposition convolution operations combined with 3x3 convolution, while reducing computational complexity, and branch 3 uses maximum pooling to achieve feature downsampling and retain key information. This unit performs multi-scale feature extraction and dimensionality reduction in deep networks;

[0109] InceptionC unit, used to adopt a multi-branch design, where branch 1 uses 1x1 convolution to extract local features, branch 2 uses 1x3 and 3x1 convolution to achieve efficient feature extraction and diversified representation, and branch 3 uses 3x3 convolution followed by 1x3 and 3x1 decomposed convolution to increase the receptive field. Through the horizontal and vertical convolution decomposition operations between branches, more complex local and global features are captured, and the ability to combine features between branches is enhanced. Convolutional attention mechanism unit, this module includes spatial attention mechanism and channel attention mechanism. This design enables the model to adaptively learn and adjust the importance of features, distinguishing key features in the channel dimension and locating important areas in the spatial dimension, so that the deep feature processing submodule can automatically adjust the importance of depth-related features and improve the model prediction accuracy;

[0110] The global average pooling unit is used to perform adaptive two-dimensional average pooling on the input feature map, and finally adjust the output feature map to the specified target size.

[0111] Specifically, compared to the traditional method of simply deepening the number of model layers, this embodiment uses the method of increasing the network width of the Inception parallel module to provide a more balanced and efficient deep learning architecture. The algorithm introduces convolution kernels of different sizes in each Inception parallel module, so that the model can extract features of different scales in parallel in the same layer, and splice the features of each branch together through feature fusion. This multi-scale feature extraction and fusion method helps the model capture various acoustic features in the marine environment, so as to better adapt to the complexity and variability of the marine environment. The Inception parallel module mainly includes an input layer and multiple convolution branches. Among them, the 1x1 convolution branch is mainly used for dimensionality reduction operations and cross-channel information integration. In addition, by decomposing the large convolution kernel by multiple small convolution kernels, the amount of parameters is reduced while maintaining the receptive field.

[0112] In order to enable the model to adjust the importance of different features, this method also introduces an attention mechanism. The entire attention mechanism process is as follows:

[0113]

[0114] Where x is the input feature, x' is the enhanced feature, M c is the channel attention module, M s is the spatial attention module, is element-wise multiplication.

[0115] The channel attention and spatial attention modules are implemented as follows:

[0116] M c (x)=sigmoid(MLP(Avgpool(x))+MLP(Maxpool(x))

[0117] The above formula represents the channel attention mechanism, which shares the parameters in MLP.

[0118] M s (x)) = sigmoid(Conv 7×7 ([Avgpool(x); Maxpool(x)]))

[0119] The above formula represents the spatial attention mechanism, Conv 7×7 Indicates a convolution operation with a convolution kernel of 7×7.

[0120] This dual attention mechanism design of CBAM allows the model to adaptively learn and adjust the importance of features, both distinguishing key features in the channel dimension and locating important areas in the spatial dimension. This approach significantly enhances the model's ability to capture complex feature interactions, thereby improving the model's accuracy and robustness in complex environments.

[0121] The hyper-parameter settings during training in this embodiment are shown in Table 1:

[0122] Table 1

[0123] Optimizer Adam Learning rate (lr) 0.001 Batch size (batch_size) 32 Epoch 60

[0124] Performance evaluation indicators:

[0125] In this embodiment, the mean absolute error (MAE) is selected as the detection index of the positioning performance, and its expression is as follows:

[0126]

[0127] Where N represents the number of test samples, y irepresents the true distance or depth value of the i-th sample, Represents the predicted distance or depth value of the i-th sample.

[0128] In this example, MAE_r (mean absolute error of distance) and MAE_d (mean absolute error of depth) are used as the main evaluation indicators of model performance. Through comprehensive comparative experiments with these two indicators, we can objectively evaluate the performance of the model in terms of prediction accuracy and systematic deviation.

[0129] Transfer Learning:

[0130] Since the model is trained only with simulated data, and the simulated environment cannot fully represent the complex ocean environment with time-varying characteristics, the generalization ability of the model is relatively weak when applied to the real environment. To solve the problem of performance degradation of the model in the real environment, the architecture cleverly uses transfer learning technology. This allows the model to transfer the knowledge learned on the simulated data to the real scene, while adapting to the specific characteristics of the real environment.

[0131] First, the basic model is trained using a simulation data set to learn the basic features in the distance and depth estimation tasks in a simulation environment. This step aims to establish a basic knowledge system, capture common features, and lay the foundation for subsequent transfer learning. Subsequently, in order to make the model more applicable to the real ocean environment, the real sea trial data is divided into training, verification, and test data in a ratio of 6:1:3. The model is fine-tuned using the training data in the real data to achieve domain migration of the model to the real environment. This process aims to retain the effective features obtained in the simulation training while gradually adjusting the model to adapt to the real environment, thereby improving the robustness and generalization ability of the model.

[0132] Table 2

[0133]

[0134] Finally, the performance of the model was verified using simulation data sets and sea trial data sets. The simulation data set was generated based on the environmental parameters collected by the SACLANT Center in the shallow waters north of Elba Island on the west coast of Italy in October 1993, and was divided into training set, verification set and test set; the sea trial mobile source data was used by the SACLANT Center on the afternoon of October 27 of the same year to conduct a mobile source experiment using an auxiliary ship to collect mobile source data, and the depth of the source buoy was marked, so this data set was used as the sea trial test set. The simulation experiment results are shown in Table 2. From the test results in the simulation data set, it can be seen that compared with the benchmark model, the average absolute error of the method proposed in this embodiment in distance and depth estimation is improved by 0.005km and 0.1m. And through the transfer learning technology, the model was also verified using the sea trial data set. The experimental results show that the average absolute error of the method in distance and depth prediction is 0.194km and 1.75m respectively. The results prove that the method has shown excellent performance in practical applications, especially in range and depth prediction.

[0135] The above are only preferred specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. An underwater passive positioning method combining multi-scale feature extraction and transfer learning, characterized in that: include: Acquire the sound wave data of the sound source to be located in the simulation environment; Extracting features from the sound wave data of the sound source to be located to obtain sound wave features of the sound source; Inputting the sound source sound wave characteristics into a positioning prediction model to predict the sound source distance and sound source depth, wherein the positioning prediction model is obtained by training a training set, and the training set is waveform data of sound pressure at different frequencies; The sound source acoustic wave localization problem in the simulation environment is domain-migrated to a real sea trial environment to perform underwater sound source localization.

2. The underwater passive positioning method combining multi-scale feature extraction and transfer learning according to claim 1, characterized in that: Acquiring the training set includes: Simulate the ocean environment and the waveform data of sound pressure at different frequencies within the preset bandwidth; Superimposing the waveform data to obtain simulation waveform data; The simulation waveform data is preprocessed to obtain the training set.

3. The underwater passive positioning method combining multi-scale feature extraction and transfer learning according to claim 2 is characterized in that: Extracting features from the sound wave data of the sound source to be located to obtain the sound wave features of the sound source includes: Performing Fourier transform on the simulation waveform data to obtain complex sound pressure data; Normalizing the complex sound pressure data using an L2 normal form to obtain a characteristic signal; A covariance operation is performed on the characteristic signal samples to obtain the sound wave characteristics of the sound source.

4. The underwater passive positioning method combining multi-scale feature extraction and transfer learning according to claim 3 is characterized in that: The method of extracting features from the sound wave data of the sound source to be located and obtaining the sound wave features of the sound source is as follows: p(f)=[p1(f),p2(f),...,p L (f)] T Where p(f) is the array signal with frequency f, p L (f) is the complex sound pressure signal of channel L at frequency f, where L is the number of vertical linear arrays of hydrophones; The method of normalizing the complex sound pressure data using the L2 norm is: in, is the normalized array signal; Then the vector obtained by the above formula is normalized for covariance processing, preliminary feature extraction is performed on the data, and the covariance matrix C is constructed. ( f ) : Where C(f) is the L×L complex covariance matrix (SCM), for The transposed conjugate matrix of .

5. The underwater passive positioning method combining multi-scale feature extraction and transfer learning according to claim 1, characterized in that: The positioning prediction model includes: an input module, a parameter sharing module, a feature processing module and an output module; The input module is used to input feature data to be located; The parameter sharing module is used for feature sharing between different tasks; The feature processing module is used to perform distance feature processing and depth feature processing; The output module is used to output the sound source distance and the sound source depth.

6. The underwater passive positioning method combining multi-scale feature extraction and transfer learning according to claim 5, characterized in that: The parameter sharing module includes: a first unit, a second unit, a convolutional attention mechanism unit and a third unit; The first unit is used to realize multi-scale feature extraction of the model; The second unit is used to reduce the spatial dimension of the feature map while retaining key features of multiple scales; The convolutional attention mechanism unit is used to distinguish key features in the channel dimension and locate important areas in the spatial dimension; The third unit is used to perform multi-scale feature extraction again based on the features processed by the convolutional attention mechanism unit to achieve feature sharing between different tasks.

7. The underwater passive positioning method combining multi-scale feature extraction and transfer learning according to claim 5, characterized in that: The feature processing module includes: a distance feature processing submodule and a depth feature submodule in parallel; The distance feature processing submodule is used to process distance features; The depth feature processing submodule is used to process depth features.

8. The underwater passive positioning method combining multi-scale feature extraction and transfer learning according to claim 7, characterized in that: The distance feature processing submodule includes: a fourth unit, a fifth unit and a global average pooling unit; The fourth unit is used for multi-scale feature extraction and dimensionality reduction in the deep network; The fifth unit is used to capture more complex local and global features and enhance the ability to combine features between branches; The global average pooling unit is used to perform adaptive two-dimensional average pooling on the processed feature map, and finally adjust the output feature map to a specified target size.

9. The underwater passive positioning method combining multi-scale feature extraction and transfer learning according to claim 7, characterized in that: The deep feature processing submodule includes a fourth unit, a fifth unit, a convolutional attention mechanism unit and a global average pooling unit; The fourth unit is used for multi-scale feature extraction and dimensionality reduction in the deep network; The fifth unit is used to capture more complex local and global features and enhance the ability to combine features between branches; The convolutional attention mechanism unit is used to distinguish key features in the channel dimension and locate important areas in the spatial dimension; The global average pooling unit is used to perform adaptive two-dimensional average pooling on the processed feature map, and finally adjust the output feature map to a specified target size.

Images