Underwater target recognition method, device, computer equipment and readable storage medium

Abstract

Embodiments of the present application provide an underwater target recognition method, apparatus, computer device, and readable storage medium. Embodiments of the present application provide an underwater target recognition method, apparatus, computer device, and readable storage medium. The method includes: obtaining an initial echo signal reflected by an underwater target for a sonar signal and generating an echo time-frequency diagram; determining at least one time-frequency sub-image corresponding to a micro-motion period according to the echo time-frequency diagram; performing projections in multiple directions on the time-frequency sub-image to obtain multiple projection intensity distributions of the time-frequency sub-image, and constructing an energy distribution image based on the multiple projection intensity distributions; obtaining a sine curve sequence corresponding to the micro-Doppler component in the time-frequency curve according to the energy distribution image; extracting the micro-Doppler component based on the sine curve sequence to generate a micro-Doppler time-frequency diagram; and inputting the micro-Doppler time-frequency diagram into a pre-trained classification model to obtain an identification result of the underwater target. In this way, the accuracy of underwater target recognition can be improved.

Description

Technical Field

[0001] The present application relates to the field of underwater acoustic detection technology, and in particular, to an underwater target recognition method, device, computer device, and readable storage medium. Background Art

[0002] The ocean covers more than 70% of the Earth's surface area, approximately 360 million square kilometers, and is an important strategic space for the sustainable development of humanity. Developing ocean resources and ocean science and technology has become a global trend. Underwater target detection technology, as an important part of ocean information perception, has played an important role in the fields of ocean science, ocean security, and ocean economy, and has great strategic value.

[0003] In related technologies, due to the good propagation performance of sound waves underwater, underwater acoustic communication can be carried out using sound waves as a carrier, and target detection can be achieved based on the extracted underwater acoustic signals. However, since the underwater acoustic channel is a complex channel with characteristics such as severe noise and multipath effect interference, and due to the acoustic energy loss caused by the absorption loss of seawater, the reflection loss of the boundary, and the spreading loss of sound waves, problems such as low signal quality and many signal interferences are caused, reducing the accuracy of underwater target recognition. Summary of the Invention

[0004] The main objective of the embodiments of the present application is to propose an underwater target recognition method, device, computer device, and readable storage medium, which can improve the accuracy of underwater target recognition.

[0005] To achieve the above objective, a first aspect of the embodiments of the present application proposes an underwater target recognition method, and the method includes:

[0006] Obtain an initial echo signal reflected by an underwater target for a sonar signal, and generate an echo time-frequency diagram based on the initial echo signal;

[0007] Determine the micro-motion period of the underwater target according to the time-frequency curve in the echo time-frequency diagram, and determine at least one time-frequency sub-image corresponding to the micro-motion period from the echo time-frequency diagram;

[0008] Perform projections of the time-frequency sub-image in multiple directions to obtain multiple projection intensity distributions of the time-frequency sub-image in different projection directions, and construct an energy distribution image based on the multiple projection intensity distributions;

[0009] Wherein, there are overlapping projections between different projection intensity distributions in the energy distribution image;

[0010] Determine a target energy region with the highest cumulative projection intensity value from the energy distribution image, and based on the target position of the target energy region in the energy distribution image, obtain a sine curve sequence corresponding to the micro-Doppler component in the time-frequency curve;

[0011] Based on the sine curve sequence, extract the micro-Doppler component from the echo time-frequency diagram, and generate a micro-Doppler time-frequency diagram according to the micro-Doppler component;

[0012] Input the micro-Doppler time-frequency diagram into a pre-trained classification model to obtain the recognition result of the underwater target.

[0013] Correspondingly, a second aspect of the embodiments of the present application proposes an underwater target recognition device, and the device includes:

[0014] An acquisition module, configured to acquire an initial echo signal reflected by an underwater target for a sonar signal, and generate an echo time-frequency diagram based on the initial echo signal;

[0015] A first determination module, configured to determine the micro-motion period of the underwater target according to the time-frequency curve in the echo time-frequency diagram, and determine at least one time-frequency sub-image corresponding to the micro-motion period from the echo time-frequency diagram;

[0016] A construction module, configured to project the time-frequency sub-image in multiple directions to obtain multiple projection intensity distributions of the time-frequency sub-image in different projection directions, and construct an energy distribution image based on the multiple projection intensity distributions; wherein, there are overlapping projections between different projection intensity distributions in the energy distribution image;

[0017] A second determination module, configured to determine a target energy region with the highest cumulative projection intensity value from the energy distribution image, and based on the target position of the target energy region in the energy distribution image, obtain a sine curve sequence corresponding to the micro-Doppler component in the time-frequency curve;

[0018] A generation module, configured to extract the micro-Doppler component from the echo time-frequency diagram based on the sine curve sequence, and generate a micro-Doppler time-frequency diagram according to the micro-Doppler component;

[0019] An input module, configured to input the micro-Doppler time-frequency diagram into a pre-trained classification model to obtain the recognition result of the underwater target.

[0020] In some embodiments, the acquisition module is further configured to:

[0021] Obtain the radial displacement velocity of the underwater target, the sound speed of the underwater environment corresponding to the underwater target, and the carrier frequency of the sonar signal, and determine the predicted Doppler frequency of the underwater target based on the radial displacement velocity, the sound speed, and the carrier frequency;

[0022] Obtain a preset filtering range coefficient, and set the upper and lower cut-off frequencies of the band-pass filter based on the carrier frequency, the filtering range coefficient, and the predicted Doppler frequency;

[0023] Filter the initial echo signal through the upper and lower cut-off frequencies to obtain a first echo signal;

[0024] Normalize the first echo signal to obtain a second echo signal;

[0025] Perform time-frequency transformation on the second echo signal to obtain the echo time-frequency diagram corresponding to the initial echo signal.

[0026] In some embodiments, the underwater target recognition device further includes a processing module, configured to:

[0027] Successively intercept the signals to be processed in the first echo signal through a preset sliding window;

[0028] For each signal to be processed corresponding to a sliding window, obtain the mean and standard deviation of the signal to be processed, and set the threshold range of the signal to be processed based on the mean and the standard deviation;

[0029] When there is a target signal value outside the threshold range in the signal to be processed, replace the target signal value with the mean;

[0030] After processing all the signals to be processed included in the first echo signal, obtain the processed first echo signal.

[0031] In some embodiments, the first determination module is further configured to:

[0032] Perform autocorrelation analysis on the echo time-frequency diagram to obtain the autocorrelation function corresponding to the initial echo signal; wherein, the autocorrelation function is used to measure the overlapping degree of the time-frequency curves at different time delays;

[0033] Obtain the first target peak of the autocorrelation function;

[0034] Determine the micro-motion period of the underwater target based on the position of the first target peak.

[0035] In some embodiments, the underwater target recognition device further includes an adjustment module, configured to:

[0036] Perform a linear transformation on the time-frequency sub-image to obtain the linear component in the time-frequency sub-image;

[0037] For each linear component, obtain the slope and intercept of each linear component, and based on the slope and the intercept, determine the second-order translational result of the linear component;

[0038] When the second-order translational result indicates that there is a second-order translation in the linear component, based on the slope and the intercept, perform a bias adjustment on the frequency-axis coordinates in the time-frequency sub-image to obtain a processed time-frequency sub-image.

[0039] In some embodiments, the underwater target recognition device further includes a training module for:

[0040] Obtain a plurality of first echo time-frequency diagrams corresponding to an underwater target in an actual environment; the first echo time-frequency diagrams correspond to various micro-motion forms;

[0041] Through a preset diffusion model, perform sample augmentation based on the plurality of first echo time-frequency diagrams to obtain corresponding second echo time-frequency diagrams;

[0042] Extract the micro-Doppler components from the first echo time-frequency diagrams and the second echo time-frequency diagrams respectively to obtain a plurality of sample micro-Doppler time-frequency diagrams, and generate a sample training set based on the plurality of sample micro-Doppler time-frequency diagrams;

[0043] Sequentially input each sample micro-Doppler time-frequency diagram in the sample training set into a preset classification model to obtain the predicted class label corresponding to each sample data;

[0044] Obtain the sample class label of the sample data, and determine the target loss based on the difference between the predicted class label and the sample class label;

[0045] Based on the target loss, adjust the parameters of the preset classification model to obtain a classification model.

[0046] In some embodiments, the training module is further configured to:

[0047] For each first echo time-frequency diagram, convert the first echo time-frequency diagram into a one-dimensional time series;

[0048] Input the one-dimensional time series into a preset diffusion model so that the diffusion model performs augmentation based on the first echo time-frequency diagram to obtain a plurality of second echo time-frequency diagrams having the same micro-motion form as the first echo time-frequency diagram; wherein, the first dimension of each convolution kernel in the diffusion model is adapted to the dimension of the one-dimensional time series.

[0049] Correspondingly, in the third aspect of the embodiments of the present application, a computer device is proposed. The computer device includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the underwater target recognition method according to any one of the embodiments of the first aspect of the present application is implemented.

[0050] Correspondingly, in the fourth aspect of the embodiments of the present application, a computer-readable storage medium is proposed. The storage medium stores a computer program, and when the computer program is executed by a processor, the underwater target recognition method according to any one of the embodiments of the first aspect of the present application is implemented.

[0051] In the embodiments of the present application, an initial echo signal reflected by an underwater target for a sonar signal is obtained, and an echo time-frequency diagram is generated based on the initial echo signal; the micro-motion period of the underwater target is determined according to the time-frequency curve in the echo time-frequency diagram, and at least one time-frequency sub-image corresponding to the micro-motion period is determined from the echo time-frequency diagram; the time-frequency sub-image is projected in multiple directions to obtain multiple projection intensity distributions of the time-frequency sub-image in different projection directions, and an energy distribution image is constructed based on the multiple projection intensity distributions; wherein, there are overlapping projections between different projection intensity distributions in the energy distribution image; a target energy region with the highest cumulative projection intensity value is determined from the energy distribution image, and based on the target position of the target energy region in the energy distribution image, a sine curve sequence corresponding to the micro-Doppler component in the time-frequency curve is obtained; based on the sine curve sequence, the micro-Doppler component is extracted from the echo time-frequency diagram, and a micro-Doppler time-frequency diagram is generated according to the micro-Doppler component; the micro-Doppler time-frequency diagram is input into a pre-trained classification model to obtain the recognition result of the underwater target. In this way, the micro-motion period of the underwater target can be determined from the echo time-frequency diagram to better understand the inherent motion characteristics of the underwater target, thereby improving the recognition accuracy. And, by using the characteristic that the main distribution directions of the micro-Doppler features are consistent and will be more concentrated in the projected space, by determining the target position of the target energy region with the highest cumulative projection intensity value in the energy distribution image, the sine curve sequence corresponding to the micro-Doppler component can be accurately determined to accurately separate the micro-Doppler component corresponding to the underwater target from the complex echo signal for the classification model to perform recognition and classification, providing clearer and more representative features for the classification model, effectively reducing the influence of interference signals, and improving the recognition accuracy. In summary, the present application can improve the accuracy of underwater target recognition. Description of the Drawings

[0052] Figure 1 is a schematic diagram of the architecture of the underwater target recognition system provided by the embodiments of the present application;

[0053] Figure 2 is a flowchart of the underwater target recognition method provided by the embodiments of the present application;

[0054] Figure 3 is the time-frequency sub-image reconstructed after the inverse Radon transform provided by the embodiments of the present application;

[0055] Figure 4 is the step diagram for extracting the micro-Doppler component provided by the embodiments of the present application;

[0056] Figure 5 is the schematic diagram of the micro-motion form provided by the embodiments of the present application;

[0057] Figure 6 is the flowchart for training the classification model provided by the embodiments of the present application;

[0058] Figure 7 is the structural diagram of the diffusion model provided by the embodiments of the present application;

[0059] Figure 8 is the schematic diagram of the functional modules of the underwater target recognition device provided by the embodiments of the present application;

[0060] Figure 9 is the schematic diagram of the hardware structure of the computer device provided by the embodiments of the present application. Detailed implementation manners

[0061] In order to make the objectives, technical solutions and advantages of the present application clearer and more understandable, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0062] It should be noted that although the functional modules are divided in the device schematic diagram and the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order from the module division in the device or the order in the flowchart. Terms such as "first" and "second" in the specification, claims and the above-mentioned drawings are used to distinguish similar objects and do not necessarily need to describe a specific order or sequence.

[0063] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present application belongs. The terms used herein are only for the purpose of describing the embodiments of the present application and are not intended to limit the present application.

[0064] The ocean covers more than 70% of the Earth's surface area, approximately 360 million square kilometers, and is an important strategic space for the sustainable development of humanity. Developing ocean resources and marine science and technology has become a global trend. Underwater target detection technology, as an important part of marine information perception, has played an important role in the fields of marine science, marine security and marine economy, and has great strategic value.

[0065] In the related art, since sound waves have good propagation performance underwater, underwater acoustic communication can be carried out by using sound waves as a carrier, and the detection of targets can be realized based on the extracted underwater acoustic signals. However, since the underwater acoustic channel is a complex channel with characteristics such as severe noise and multipath effect interference, and due to the acoustic energy loss caused by the absorption loss of seawater, the reflection loss of the boundary, and the spreading loss of sound waves, problems such as low signal quality and many signal interferences occur, reducing the accuracy of underwater target recognition.

[0066] Based on this, the embodiments of the present application provide an underwater target recognition method, device, computer device, and readable storage medium, which can improve the accuracy of underwater target recognition.

[0067] The underwater target recognition method, device, computer device, and readable storage medium provided by the embodiments of the present application will be specifically described through the following embodiments. First, the underwater target recognition system in the embodiments of the present application will be described.

[0068] Please refer to Figure 1 , in some embodiments, the embodiments of the present application provide an underwater target recognition system, including a terminal 11 and a server side 12.

[0069] Exemplarily, the terminal 11 can be a sonar sensor, an autonomous underwater vehicle, a data acquisition system, etc. The terminal 11 can interact with the underwater environment, be responsible for sending sonar signals, collecting the initial echo signals reflected by the underwater target for the sonar signals, and performing preliminary processing on the initial echo signals, such as filtering, processing of outliers, etc.

[0070] Furthermore, the server side 12 can be a high-performance computing server, a cloud server, a data center, or other computer devices. The server side 12 can receive the initial echo signals sent by the terminal 11, perform processing on the initial echo signals such as time-frequency analysis and feature extraction, and run a classification model to identify underwater targets.

[0071] The underwater target recognition method in the embodiments of the present application will be described through the following embodiments.

[0072] It should be noted that in each specific embodiment of this application, when it comes to performing relevant processing based on data related to the user's identity or characteristics, such as user information, user behavior data, user historical data, and user location information, the user's permission or consent will be obtained first. Moreover, the collection, use, and processing of these data will comply with relevant laws, regulations, and standards. In addition, when the embodiments of this application need to obtain the user's sensitive personal information, the user's separate permission or separate consent will be obtained through methods such as pop-up windows or redirecting to a confirmation page. After clearly obtaining the user's separate permission or separate consent, the necessary user-related data for the normal operation of the embodiments of this application will be obtained.

[0073] In the embodiments of this application, a description will be made from the dimension of an underwater target recognition device, which can be specifically integrated in a computer device. Refer to Figure 2 , Figure 2 which is the flowchart of the steps of the underwater target recognition method provided by the embodiments of this application. Taking the example that the underwater target recognition device is specifically integrated in a terminal or a server, when the processor on the terminal or the server executes the program instructions corresponding to the underwater target recognition method, the specific process is as follows:

[0074] Step 101, obtain the initial echo signal reflected by the underwater target for the sonar signal, and generate an echo time-frequency diagram based on the initial echo signal.

[0075] In some embodiments, to facilitate the classification and recognition of underwater targets, a sonar signal can be sent to the underwater target. After receiving the initial echo signal reflected by the underwater target, the complex initial echo signal (time-domain signal) is converted into an intuitive time-frequency domain representation to provide a basis for the accurate classification and recognition of underwater targets.

[0076] Among them, the underwater target can be various objects or entities in the water environment, such as natural targets like marine organisms, and man-made targets such as ships, submarines, torpedoes, underwater robots, and underwater facilities (such as engines, rotor blades, etc.). Different underwater targets have their unique micro-motion forms (such as vibration, rotation, tumbling, coning), and these micro-motion forms will form unique micro-Doppler features in the echo signal to facilitate the recognition of underwater targets.

[0077] Among them, the sonar signal can be an acoustic wave signal emitted by a sonar device for detecting underwater targets. The acoustic wave signal can be represented in the form of a sine wave, for example , where is the amplitude, is the carrier frequency of the sonar signal, and t is the time.

[0078] Among them, the initial echo signal can be the original sonar signal reflected by an underwater target, which contains various information about the underwater target, including but not limited to the physical characteristics and dynamic behavior information of the underwater target, such as the size, shape, material, and motion state of the underwater target. The initial echo signal can be expressed as , where is the amplitude of the reflected signal, is the carrier frequency of the sonar signal, is the Doppler frequency of the initial echo signal, and there is , where is the Doppler frequency generated by the translational motion of the underwater target, is the micro-Doppler frequency generated by a certain part of the underwater target.

[0079] Among them, the echo time-frequency diagram can be a two-dimensional image generated by performing time-frequency analysis on the initial echo signal through time-frequency transformation (such as short-time Fourier transform), which shows the frequency components of the signal at different time points and can facilitate more intuitive observation and analysis of micro-Doppler characteristics.

[0080] In some embodiments, a sonar device can be used to transmit a sonar signal to an underwater target and collect the initial echo signal reflected from the underwater target. After receiving the initial echo signal, the received echo signal can be preliminarily processed, including band-pass filtering, normalization, etc., to remove noise and interference. Further, short-time Fourier transform can be used to perform time-frequency analysis on the preprocessed echo signal to obtain the echo time-frequency diagram.

[0081] By obtaining the initial echo signal and generating the corresponding echo time-frequency diagram, the characteristics of the underwater target can be effectively extracted and analyzed, providing an important basis for the subsequent identification and classification of the underwater target.

[0082] In some embodiments, in order to improve the accuracy and reliability of the analysis, the initial echo signal can be subjected to band-pass filtering and normalization processing to effectively suppress noise, and at the same time, the amplitudes of all signals are scaled to the same scale for easy comparison and analysis, optimizing the use of computing resources. For example, "generating an echo time-frequency diagram based on the initial echo signal" in step 101 can include:

[0083] (101.1) Obtain the radial displacement velocity of the underwater target, the sound speed of the underwater environment corresponding to the underwater target, and the carrier frequency of the sonar signal, and determine the predicted Doppler frequency of the underwater target based on the radial displacement velocity, sound speed, and carrier frequency;

[0084] (101.2) Obtain the preset filtering range coefficient, and set the upper and lower cut-off frequencies of the band-pass filter based on the carrier frequency, filtering range coefficient, and predicted Doppler frequency;

[0085] (101.3) Filter the initial echo signal through the upper and lower cut-off frequencies to obtain the first echo signal;

[0086] (101.4) Normalize the first echo signal to obtain the second echo signal;

[0087] (101.5) Perform time-frequency transformation on the second echo signal to obtain the echo time-frequency diagram corresponding to the initial echo signal.

[0088] Among them, the radial displacement velocity can be the velocity component of the underwater target relative to the sonar device along the sound wave propagation direction, which determines the magnitude of the Doppler frequency shift.

[0089] Among them, the sound speed can be the propagation speed of sound waves in the underwater environment.

[0090] Among them, the carrier frequency can be the basic frequency of the sound wave signal emitted by the sonar device.

[0091] Among them, the predicted Doppler frequency can be the range of the change in the echo signal frequency caused by the movement of the target.

[0092] Among them, the filtering range coefficient can be a proportionality factor used to determine the upper and lower cut-off frequencies of the band-pass filter, which determines the bandwidth range of the filter and can be set according to the actual situation.

[0093] Among them, the upper and lower cut-off frequencies can be the lowest and highest frequencies allowed to pass through the band-pass filter.

[0094] Among them, the first echo signal can be the initial echo signal after band-pass filtering, removing low-frequency noise and other interference signals.

[0095] Among them, the second echo signal can be the result of normalizing the first echo signal, and the normalization operation can make the signal amplitude unified to a specific range (for example, -0.5 to 0.5).

[0096] Exemplarily, the predicted Doppler frequency can be calculated by the ratio of twice the product of the radial displacement velocity and the carrier frequency to the sound speed. For example, the specific calculation formula can be , where represents the radial displacement velocity, represents the carrier frequency of the sonar signal, represents the sound speed.

[0097] In some embodiments, the initial echo signal can be input into a band-pass filter with designed upper and lower cut-off frequencies for processing, so as to retain the components within a specific frequency range in the initial echo signal and remove the noise or interference of other frequencies, thereby obtaining a clearer and more target-feature-focused first echo signal, which is convenient for subsequent analysis and processing. Exemplarily, the upper and lower cut-off frequencies of the band-pass filter can be set as . Among them, represents the carrier frequency of the sonar signal, is the filtering range coefficient, represents the estimated Doppler frequency.

[0098] In some embodiments, the cut-off frequency of the band-pass filter can be adjusted according to the actual signal-to-noise ratio of the sonar signal. For example, when high-frequency noise (greater than the preset frequency threshold) is detected, the filtering range coefficient α is automatically reduced to suppress interference; in a low-noise (less than the preset noise threshold) environment, α is expanded to retain more effective signals, and the scale of the adjusted filtering range coefficient α can be set according to the actual situation, which is not limited in the embodiments of the present application.

[0099] Furthermore, in order to unify the feature scale, the first echo signal can be normalized to ensure the accuracy of the calculation. Exemplarily, the normalization formula can be: . Among them, can be the result after normalization, can be a single data point in the initial echo signal to be normalized, is the maximum value in the initial echo signal, is the minimum value in the initial echo signal. By normalizing each data point in the first echo signal, the second echo signal after processing the first echo signal can be obtained.

[0100] Exemplarily, through time-frequency transformation (such as short-time Fourier transform), the second echo signal can be transformed from the time domain to the time-frequency domain to generate an intuitive two-dimensional image, that is, the echo time-frequency diagram. The horizontal axis in the echo time-frequency diagram can represent time (seconds), and the vertical axis represents frequency (hertz).

[0101] Through the above methods, the noise and interference in the signal can be effectively filtered out, and the key frequency part of the signal can be retained; in addition, the normalization processing ensures the consistency of the signal feature scale and improves the accuracy of subsequent analysis. Finally, the time-frequency transformation converts the processed second echo signal into an echo time-frequency diagram, providing an intuitive two-dimensional image for further analysis and underwater target recognition.

[0102] In some embodiments, in order to improve the overall stability of the signal and avoid distortion of subsequent processing results caused by individual outliers, outliers in the first echo signal can be identified and processed to effectively smooth the signal, reduce the influence of mutation values, and improve the quality of the signal. For example, before (101.4), that is, before "performing normalization processing on the first echo signal to obtain a second echo signal", the following may also be included:

[0103] (A.1) Sequentially intercept the signal to be processed in the first echo signal through a preset sliding window;

[0104] (A.2) For the signal to be processed corresponding to each sliding window, obtain the mean and standard deviation of the signal to be processed, and set the threshold range of the signal to be processed based on the mean and standard deviation;

[0105] (A.3) When there is a target signal value outside the threshold range in the signal to be processed, replace the target signal value with the mean;

[0106] (A.4) After processing all the signals to be processed included in the first echo signal, obtain the processed first echo signal.

[0107] Among them, the sliding window can be a time window with a fixed length. The sliding window can slide from beginning to end on the signal, and each time a signal segment with a fixed length is intercepted. The size and step length of the sliding window can be adjusted according to specific requirements.

[0108] Among them, the signal to be processed can be each segment of the first echo signal intercepted in the sliding window. These segments are the basis for subsequent mean, standard deviation calculation, and threshold range setting.

[0109] Among them, the standard deviation can be a statistic used to measure the degree of dispersion of the signal to be processed, indicating the degree of deviation of the data from its mean.

[0110] Among them, the threshold range can be a range interval used to identify outliers. Signal values exceeding the threshold range can be determined as outliers.

[0111] Among them, the target signal value can be the signal value in the signal to be processed that is outside the corresponding threshold range.

[0112] Exemplarily, the threshold range can be , where μ is the mean of multiple signal values in the signal to be processed, σ is the standard deviation of multiple signal values in the signal to be processed, k is a preset coefficient (for example, k = 3 represents three times the standard deviation, and the value of k can be set according to the actual situation), and the target signal value exceeding the threshold range is regarded as an outlier or a mutation point.

[0113] In certain embodiments, the first echo signal can be successively segmented into multiple smaller segments through a sliding window to obtain multiple signals to be processed intercepted by the sliding window. Subsequently, for each signal to be processed, the mean and standard deviation of all signal values included in the signal to be processed can be calculated, and a corresponding threshold range for the signal to be processed can be set based on the mean and standard deviation. For example, after calculating the mean μ and standard deviation σ for the signal to be processed 1, if k is 3, the threshold range for the signal to be processed 1 can be determined as .

[0114] In some embodiments, signal values outside the threshold range can be replaced with the mean, thereby effectively smoothing the signal and reducing the influence of mutation values.

[0115] Exemplarily, in addition to determining the threshold range through the mean and standard deviation, the threshold range can also be determined through the percentile method, the median absolute deviation (calculating the absolute deviation of each data point from the median and using a certain percentile of the deviation as the value), and so on.

[0116] Exemplarily, the first echo signal can also not be divided, and the overall mean and overall standard deviation of all signal values in the first echo signal can be directly obtained, and the threshold range can be determined based on the overall mean and overall standard deviation. When there are signal values outside the threshold range in the first echo signal, the signal value is replaced with the overall mean. In this way, the efficiency of processing outliers in the first echo signal can be improved.

[0117] By dividing the first echo signal into multiple signals to be processed, the signal can be processed within smaller local signal segments to more accurately identify and process local features, so as to adapt to the signal characteristics and application requirements at different stages and improve the accuracy and robustness of signal processing.

[0118] Step 102: Determine the micro-motion period of the underwater target according to the time-frequency curve in the echo time-frequency diagram, and determine at least one time-frequency sub-image corresponding to the micro-motion period from the echo time-frequency diagram.

[0119] In some embodiments, in order to centrally process the key features of the target and reduce the interference of irrelevant information, the micro-motion period of the underwater target can be determined, and the time-frequency sub-image corresponding to the micro-motion period can be determined to provide a basis for subsequent micro-Doppler feature extraction.

[0120] Among them, the time-frequency curve can be the trajectory representing the change of frequency components over time in the echo time-frequency diagram, which is used to display the frequency component distribution of the signal at different time points. The time-frequency curve can be generated from the initial echo signal through time-frequency transformation (such as short-time Fourier transform).

[0121] Among them, the micro-motion period can be the time required for a complete cycle of a specific micro-motion form (such as vibration, rotation, etc.) of an underwater target. For example, for simple harmonic vibration, the micro-motion period can be determined through autocorrelation analysis or spectral analysis. The micro-motion period is a key parameter for identifying and extracting micro-Doppler features.

[0122] Among them, the time-frequency sub-image can be a small part of the image intercepted from the complete echo time-frequency diagram, usually corresponding to one or more complete micro-motion periods. The time-frequency sub-image contains the main information of the target micro-motion characteristics, which is convenient for further processing and analysis.

[0123] In some embodiments, autocorrelation analysis can be performed on the time-frequency curve in the echo time-frequency diagram to estimate its micro-motion period; alternatively, the periodic characteristics of the time-frequency curve can be determined through spectral analysis. For example, by finding the significant peak positions in the spectrogram, the micro-motion period can be determined. Then, based on the results of autocorrelation analysis or spectral analysis, the micro-motion period of the underwater target is estimated.

[0124] Furthermore, since the Doppler effect caused by micro-motion is manifested as periodic characteristics in the time-frequency diagram, a complete micro-motion period contains all the key information of this effect, such as the mode and amplitude of frequency change, etc. Therefore, in order to reduce the complexity and computational amount of data processing and avoid redundancy and confusion during feature extraction, one or more target time periods corresponding to complete micro-motion periods can be intercepted from the echo time-frequency diagram, and the time-frequency sub-image corresponding to the target time period is intercepted from the echo time-frequency diagram. Each time-frequency sub-image should contain a complete micro-motion period for subsequent processing.

[0125] By determining the micro-motion period of the underwater target and determining the time-frequency sub-image based on the micro-motion period, the analysis of a single micro-motion period can be concentrated to improve the accuracy of feature extraction and more precisely capture the micro-motion characteristics without being interfered by other periods.

[0126] In some embodiments, in order to provide an accurate time reference for subsequent micro-Doppler feature extraction and classification recognition, autocorrelation analysis can be performed on the initial echo signal to facilitate the accurate identification of the periodic characteristics in the initial echo signal and facilitate the identification of the micro-motion period of the underwater target. For example, "determining the micro-motion period of the underwater target according to the time-frequency curve in the echo time-frequency diagram" in step 102 can include:

[0127] (102.1) Perform autocorrelation analysis on the echo time-frequency diagram to obtain the autocorrelation function corresponding to the initial echo signal; among them, the autocorrelation function is used to measure the overlapping degree of the time-frequency curve at different time delays;

[0128] (102.2) Obtain the first target peak of the autocorrelation function;

[0129] (102.3) Determine the micro-motion period of the underwater target based on the position of the first target peak.

[0130] Among them, the autocorrelation function can be used to measure the similarity or overlap degree of the signal at different time delays.

[0131] Among them, the first target peak can be the peak position in the autocorrelation function that is significantly higher than the background noise level for the first time. It corresponds to the main periodic component of the signal, that is, the micro-motion period of the target, and is the key reference point for determining the micro-motion period.

[0132] Furthermore, assume that the time-frequency curve is , and its autocorrelation function can be expressed as:

[0133] ;

[0134] Among them, is the time delay. By calculating at different time delays , the autocorrelation function can be obtained.

[0135] Furthermore, since the first target peak of the autocorrelation function often corresponds to the periodic characteristics of the time-frequency curve, therefore, by analyzing the graph of the autocorrelation function, the position of the first target peak in the autocorrelation function can be determined (excluding the significant peaks other than ). The first target peak indicates that the time-frequency curve is most similar to itself at this time delay. The position of the first target peak corresponds to the periodic characteristics of the time-frequency curve. Therefore, it can be used as the micro-motion period of the underwater target.

[0136] Exemplarily, if the position of the first target peak is 0.5 seconds, then the micro-motion period of the underwater target is 0.5 seconds, which means that the micro-motion characteristics of the underwater target repeat every 0.5 seconds.

[0137] In some embodiments, continuous wavelet transform can also be performed on the echo time-frequency diagram to extract the local period of the time-frequency diagram, and weighted fusion is performed in combination with the micro-motion period obtained by global analysis of the autocorrelation function to determine the final micro-motion period. For the micro-motion period result, the local period can correspond to the first period weight, and the micro-motion period obtained by autocorrelation function analysis can correspond to the second period weight. The first period weight and the second period weight can be adjusted based on the noise intensity. For example, when the noise is strong, increase the first period weight. The specific adjustment scale can be set according to the actual situation.

[0138] By performing autocorrelation analysis on the echo time-frequency diagram to determine the micro-motion period of the underwater target, it is convenient to effectively extract the micro-motion characteristics of the target, providing important information for subsequent target recognition and classification.

[0139] Step 103: Project the time-frequency sub-image in multiple directions to obtain multiple projection intensity distributions of the time-frequency sub-image in different projection directions, and construct an energy distribution image based on the multiple projection intensity distributions; wherein, there are overlapping projections between different projection intensity distributions in the energy distribution image.

[0140] In some embodiments, to improve the accuracy of underwater target recognition, by performing multi-directional projection on the time-frequency sub-image, the micro-motion characteristics of the target can be captured from different angles, thereby obtaining more comprehensive information to enhance the final target classification and recognition effect.

[0141] Among them, the projection intensity distribution can be obtained by linearly integrating or accumulating the time-frequency sub-image along each direction (such as 0°, 45°, 90°, 135°, etc.) to obtain a one-dimensional intensity distribution in that direction. Specifically, for each projection direction, all pixel values in the time-frequency sub-image are accumulated along that direction to generate a one-dimensional array representing the intensity distribution in that direction.

[0142] Among them, the energy distribution image can be a two-dimensional image constructed based on multiple projection intensity distributions in different directions.

[0143] In some embodiments, the time-frequency sub-image can be first binarized to obtain the binarized time-frequency sub-image, and then the binarized time-frequency sub-image is projected in multiple directions. In this way, the image can be simplified, the target area can be highlighted, noise interference can be reduced, and the efficiency and accuracy of subsequent processing can be improved. Exemplarily, when binarizing the time-frequency sub-image, each pixel point in the image can be converted into black and white according to a certain threshold, and the threshold can be set according to the actual situation. When the pixel value is greater than the threshold, it is set to white (such as 1 or 255 can be used); when the pixel value is less than the threshold, it is set to black (such as 0 can be used), and so on.

[0144] Exemplarily, multiple directions can include 0°, 45°, 90°, 135°, etc., and the selected directions should cover the possible micro-motion directions of the underwater target, and can be specifically selected according to the actual situation.

[0145] Furthermore, the time-frequency sub-image can be applied with the Radon transform, that is, the time-frequency sub-image is projected along multiple directions, and the time-frequency data is projected into the Radon domain, and each projection corresponds to a specific angle. Exemplarily, the formula of the Radon transform can be as follows:

[0146] ;

[0147] Among them, is the projection intensity in the Radon domain, is the time-frequency sub-image, is the distance from the origin to the straight line, is the direction of the projection.

[0148] Furthermore, for each projection direction, the projection intensity distribution of the Radon transform result that can be calculated reflects the concentration degree of the signal energy in this direction. Exemplarily, the projection intensity distribution can be calculated by the following formula:

[0149] ;

[0150] wherein, is the projection intensity distribution in the direction above.

[0151] Furthermore, after obtaining the projection intensity distributions of all directions, the projection intensity distributions of all directions can be integrated into a two-dimensional image to form an energy distribution image. For example, a two-dimensional array can be created with the same size as the Radon transform result. For each direction θ, the corresponding projection intensity distribution is filled into a row or a column of the energy distribution image. At the same time, color coding can be used to represent different energy levels. Usually, high-energy regions are represented by bright colors, and low-energy regions are represented by dark colors. After filling, the energy distribution image can be obtained.

[0152] By constructing the energy distribution image, the accuracy and reliability of underwater target recognition can be significantly improved, facilitating the subsequent accurate recognition of micro-Doppler features.

[0153] In some embodiments, in order to improve the accuracy and efficiency of signal processing, the straight-line components in the time-frequency sub-image can be identified. When there is a second-order translation in the straight-line components and corresponding translational compensation is performed, after correcting the frequency shift effect caused by the translation, projection processing can be performed to avoid masking or distorting the true micro-Doppler features. Exemplarily, before step 103, that is, before "performing projections of the time-frequency sub-image in multiple directions to obtain multiple projection intensity distributions of the time-frequency sub-image in different projection directions and constructing an energy distribution image based on the multiple projection intensity distributions", it may further include:

[0154] (B.1) Performing a linear transformation on the time-frequency sub-image to obtain the straight-line components in the time-frequency sub-image;

[0155] (B.2) For each straight-line component, obtaining the slope and intercept of each straight-line component, and determining the second-order translation result of the straight-line component based on the slope and intercept;

[0156] (B.3) When the second-order translational result indicates that there is second-order translation in the linear component, based on the slope and intercept, the frequency-axis coordinates in the time-frequency sub-image are offset adjusted to obtain the processed time-frequency sub-image.

[0157] Among them, the linear component can be the micro-Doppler effect caused by the translation of an underwater target (such as uniform forward or backward movement, etc.), which appears as a straight line at a certain angle to the time axis in the time-frequency sub-image.

[0158] Among them, the slope can be the degree of inclination of the linear component in the time-frequency sub-image, that is, the rate of change of frequency with time.

[0159] Among them, the intercept can be the intersection point of the linear component and the frequency axis in the time-frequency sub-image, representing the initial frequency of the underwater target at the moment of time t = 0, that is, the intercept of the linear component is associated with the initial Doppler frequency of the underwater target.

[0160] Among them, the second-order translational result can be the analysis result of determining whether there is second-order translation (i.e., acceleration) by analyzing the linear component. If there is second-order translation (i.e., the slope changes with time), it indicates that the underwater target not only has translational velocity but also acceleration.

[0161] Among them, the frequency-axis coordinates can be the coordinate axis representing frequency in the time-frequency sub-image.

[0162] Exemplarily, if the time-frequency sub-image shows a straight line due to the uniform linear motion of the underwater target in the time-frequency sub-image, the Radon transform or other linear transformation methods can be used to extract this linear component. If the slope of the identified linear component is 2 Hz / s and the intercept is 100 Hz (i.e., 100 hertz), this indicates that the underwater target is moving away from the sonar device at a speed of 2 Hz / s (i.e., 2 hertz per second), and the initial Doppler frequency is 100 Hz.

[0163] Furthermore, if it is detected that the slope increases with time, it indicates that the underwater target is accelerating away from the sonar device. The frequency-axis coordinates can be adjusted by subtracting the slope on the frequency axis to obtain the processed time-frequency sub-image. For example, if the slope is 2 Hz / s, 2 Hz can be subtracted on the frequency axis to compensate for this acceleration effect.

[0164] Through the above steps, the translational components of the target can be effectively extracted and analyzed from the time-frequency sub-image, and the Doppler frequency shift caused by translation can be compensated, thereby improving the accuracy and reliability of underwater target recognition.

[0165] Step 104: Determine the target energy region with the highest cumulative projection intensity from the energy distribution image, and based on the target position of the target energy region in the energy distribution image, obtain the sine curve sequence corresponding to the micro-Doppler component in the time-frequency curve.

[0166] In some embodiments, since the signal features that are significant in multiple projection directions correspond to the micro-motion behavior of the underwater target, the micro-Doppler component in the time-frequency sub-image can be effectively identified by determining the target energy region with the highest cumulative projection intensity, and the sine curve sequence can be extracted from the target energy region, further enhancing the feature description ability and improving the accuracy of underwater target recognition.

[0167] Among them, the cumulative projection intensity can be the cumulative intensity of the overlapping part between the projection intensity distributions in different directions in the energy distribution image. By calculating the sum of the intensities of these overlapping parts, the most significant target energy region in the energy distribution image can be found.

[0168] Among them, the target energy region can be a local region with the highest cumulative projection intensity in the energy distribution image, and the target energy region corresponds to the main micro-Doppler feature in the time-frequency sub-image.

[0169] Among them, the target position can be the most important coordinate position of the target energy region in the energy distribution image, which can help locate the key micro-Doppler component in the time-frequency curve.

[0170] Among them, the sine curve sequence can be the sine waveform corresponding to the target micro-motion form extracted from the time-frequency curve. The sine curve sequence reflects the micro-motion characteristics of the target (such as vibration, rotation, etc.), and is the key feature for subsequent classification and recognition.

[0171] Please refer to Figure 3 , in some embodiments, after performing projections in multiple directions on the time-frequency sub-image (converting the time-frequency sub-image from the spatial domain to the Radon domain, that is, the projection domain) to obtain the energy distribution image, the inverse Radon transform can be performed on the energy distribution image, that is, converting the data back to the spatial domain to reconstruct the time-frequency sub-image, and in the reconstructed time-frequency sub-image, identify the target energy region with the largest cumulative projection intensity, as Figure 3 shown, Figure 3 is a simplified diagram of the reconstructed time-frequency sub-image, Figure 3 in which the target energy region is marked. The target energy region corresponds to the micro-Doppler component in the echo time-frequency diagram. Alternatively, the reconstructed video sub-image can also be a black background, and the target energy region is highlighted in white.

[0172] In some embodiments, since the target energy region with the maximum cumulative projection intensity value in the reconstructed time-frequency sub-image appears as a brighter region, the region with the highest pixel brightness in the reconstructed time-frequency sub-image can be directly identified as the target energy region.

[0173] Exemplarily, for the reconstructed time-frequency sub-image, each pixel region (i.e., each pixel point) can be traversed, and the corresponding cumulative projection intensity value can be determined for each pixel region. Then, the cumulative projection intensity values of all pixel regions are compared to determine the target energy region with the highest cumulative projection intensity value, and the target energy region corresponds to the characteristics of the micro-Doppler component.

[0174] In some embodiments, when the target energy region contains multiple pixel points, when determining the target position, the centroid coordinates of the entire target energy region can be calculated using the centroid method as the target position, or the position of the pixel point with the highest cumulative projection intensity value in the target energy region can be further calculated as the target position, etc.

[0175] Exemplarily, if there are multiple high-energy regions, multiple local maximum points can be selected, corresponding to different micro-Doppler components respectively. Specifically, an energy threshold can be set, and all cumulative projection intensity values are compared with the energy threshold, and the high-energy regions corresponding to the cumulative projection intensity values higher than the energy threshold can also be determined as the target energy regions.

[0176] Furthermore, since the micro-Doppler component appears as a sine curve form in the time-frequency sub-image. Therefore, based on the position of the target energy region, the sine curve sequence corresponding to the micro-Doppler component can be determined. Exemplarily, according to the target position , the amplitude and initial phase of the sine curve can be calculated. Specifically, the amplitude is calculated as follows:

[0177] ;

[0178] Specifically, the initial phase is calculated as follows:

[0179] ;

[0180] Furthermore, based on the amplitude and initial phase of the sine curve, the sine curve sequence can be determined:

[0181] ;

[0182] where f is the frequency of the micro-Doppler component, is the initial phase, is the amplitude, and t is the time variable.

[0183] By calculating the cumulative intensity value of the overlapping regions containing the detailed micro-Doppler information generated by the vibration or rotation of the target in multiple directions, the target energy region representing the characteristics of the micro-Doppler components can be quickly located and identified. This not only speeds up the calculation but also enables the processing of complex high-energy regions, ensuring the accurate extraction of the micro-Doppler components in the form of sine curves from the time-frequency sub-image and providing reliable data support for subsequent analysis.

[0184] Step 105: Based on the sine curve sequence, extract the micro-Doppler components from the echo time-frequency diagram and generate a micro-Doppler time-frequency diagram according to the micro-Doppler components.

[0185] In some embodiments, to provide high-quality data support for subsequent classification and recognition tasks, the micro-Doppler components can be accurately extracted from the echo time-frequency diagram based on the sine curve sequence, focusing on the micro-Doppler components, reducing the interference of other irrelevant information, and improving the accuracy and robustness of underwater target recognition.

[0186] Among them, the micro-Doppler components can be the frequency changes generated by the micro-motion of the local parts (such as vibration, rotation, etc.) of the underwater target during its movement. These frequency changes usually appear as specific patterns in the echo time-frequency diagram, different from the main translational frequency (Doppler frequency) of the target. Therefore, to accurately identify the underwater target, the micro-Doppler components can be extracted from the echo time-frequency diagram to avoid the interference of irrelevant signals.

[0187] Among them, the micro-Doppler time-frequency diagram can be the part of the echo time-frequency diagram that only contains the micro-Doppler components, which shows the distribution of the micro-motion characteristics of the underwater target in time and frequency, removing the main translational frequency components and other noise interferences.

[0188] In some embodiments, the generated sine curve sequence can be mixed with the original time-frequency curve to facilitate highlighting the sine curve. After mixing, the micro-Doppler components shifted to the baseband after mixing can be effectively separated by a low-pass filter with an appropriate cut-off frequency, while suppressing other interference signals.

[0189] In some embodiments, to ensure that all micro-Doppler components are completely extracted, after each extraction of the micro-Doppler components (i.e., sine curves), the energy of the remaining signal can be calculated. If the remaining energy is lower than a preset threshold, it indicates that all significant micro-Doppler components have been extracted. Otherwise, the extraction of the micro-Doppler components needs to continue. Alternatively, the change rate of the sine curve parameters (such as frequency, amplitude, phase) in two consecutive iterations can also be calculated. If the change rate is less than the change rate threshold, it may indicate that the last curve has been reached and no further extraction is required. Otherwise, the extraction of the micro-Doppler components needs to continue.

[0190] Please refer to Figure 4 , in some embodiments, when it is necessary to continue extracting the micro-Doppler components, the above steps can be repeated. That is, after extracting the micro-Doppler components from the echo time-frequency diagram, repeat determining the micro-motion period of the underwater target according to the time-frequency curve in the updated echo time-frequency diagram, and determining at least one time-frequency sub-image corresponding to the micro-motion period from the echo time-frequency diagram. Project the time-frequency sub-image in multiple directions to obtain multiple projection intensity distributions of the time-frequency sub-image in different projection directions, and construct an energy distribution image based on the multiple projection intensity distributions. Determine the target energy region with the highest cumulative projection intensity value from the energy distribution image, and based on the target position of the target energy region in the energy distribution image, obtain the sine curve sequence corresponding to the micro-Doppler components in the time-frequency curve. Based on the sine curve sequence, extract the micro-Doppler components from the echo time-frequency diagram, and generate a micro-Doppler time-frequency diagram according to the micro-Doppler components until the extraction of the micro-Doppler components is completed. It should be noted that the implementation details of this part have been elaborated above and will not be repeated here.

[0191] Furthermore, the extracted micro-Doppler components can be mapped to the time-frequency domain to generate a micro-Doppler time-frequency diagram, where the horizontal axis can represent time, the vertical axis represents frequency, and the color or brightness represents the intensity or energy of the signal.

[0192] By extracting the micro-Doppler components and generating the micro-Doppler time-frequency diagram in the above manner, the inverse Radon transform mainly applied in the imaging field can be applied to the extraction of the micro-Doppler characteristics and target recognition of underwater targets. Different from the recognition of underwater targets through simple signal processing (such as filtering), the present application can effectively utilize the characteristics after the projection of the micro-Doppler components, accurately extract the micro-Doppler components from the echo time-frequency diagram, making the extraction of the micro-Doppler components more targeted and accurate, having wide engineering practical value, and promoting the development of underwater target detection technology.

[0193] Step 106, input the micro-Doppler time-frequency diagram into a pre-trained classification model to obtain the recognition result of the underwater target.

[0194] In some embodiments, in order to accurately identify the type and micro-motion form of the underwater target, a classification model trained with a large amount of data can be used to classify and identify the micro-Doppler time-frequency diagram to improve the recognition efficiency and reliability.

[0195] Among them, the classification model can be a machine learning or deep learning model for mapping the micro-Doppler time-frequency diagram to predefined class labels. The classification model is trained with a large number of sample micro-Doppler time-frequency diagrams of different classes and has good classification ability.

[0196] Among them, the recognition result can be the predicted class label output by the classification model after processing the micro-Doppler time-frequency diagram, which represents the micro-motion form that the component model believes the input data corresponds to.

[0197] Exemplarily, the classification model can be a ResNet-18 model, or other models in the ResNet series, or other deep learning models. The embodiments of the present application do not make specific limitations thereto.

[0198] Please refer to Figure 5 , specifically, the recognition results of underwater targets can include four micro-motion forms: vibration, rotation, tumbling, and coning, or can also be adjusted according to the actual classification model and training parameters.

[0199] In the embodiments of the present application, the initial echo signal reflected by the underwater target for the sonar signal is obtained, and the echo time-frequency diagram is generated based on the initial echo signal; the micro-motion period of the underwater target is determined according to the time-frequency curve in the echo time-frequency diagram, and at least one time-frequency sub-image corresponding to the micro-motion period is determined from the echo time-frequency diagram; the time-frequency sub-image is projected in multiple directions to obtain multiple projection intensity distributions of the time-frequency sub-image in different projection directions, and an energy distribution image is constructed based on the multiple projection intensity distributions; among them, there are overlapping projections between different projection intensity distributions in the energy distribution image; the target energy region with the highest cumulative projection intensity value is determined from the energy distribution image, and based on the target position of the target energy region in the energy distribution image, the sine curve sequence corresponding to the micro-Doppler component in the time-frequency curve is obtained; based on the sine curve sequence, the micro-Doppler component is extracted from the echo time-frequency diagram, and the micro-Doppler time-frequency diagram is generated according to the micro-Doppler component; the micro-Doppler time-frequency diagram is input into the pre-trained classification model to obtain the recognition result of the underwater target. In this way, the micro-motion period of the underwater target can be determined from the echo time-frequency diagram to better understand the inherent motion characteristics of the underwater target, thereby improving the recognition accuracy. And, the main distribution directions of the micro-Doppler features are consistent, and they will appear more concentrated in the projected space. By determining the target position of the target energy region with the highest cumulative projection intensity value in the energy distribution image, the sine curve sequence corresponding to the micro-Doppler component can be accurately determined to accurately separate the micro-Doppler component corresponding to the underwater target from the complex echo signal for the classification model to perform recognition and classification, providing clearer and more representative features for the classification model, effectively reducing the influence of interference signals, and improving the recognition accuracy. In summary, the present application can improve the accuracy of underwater target recognition.

[0200] In some embodiments, in order to enable the classification model to more accurately predict the sample class label and improve the classification accuracy, the preset classification model can be iteratively trained multiple times to gradually reduce the target loss, and finally a classification model with excellent performance and higher classification accuracy can be obtained. Exemplarily, the classification model can be trained in the following manner:

[0201] (C.1) Obtain multiple first echo time-frequency diagrams corresponding to an underwater target in the actual environment; the first echo time-frequency diagrams correspond to various micro-motion forms;

[0202] (C.2) Based on the preset diffusion model, perform sample augmentation on the basis of multiple first echo time-frequency diagrams to obtain corresponding multiple second echo time-frequency diagrams;

[0203] (C.3) Extract the micro-Doppler components from the first echo time-frequency diagrams and the second echo time-frequency diagrams respectively to obtain multiple sample micro-Doppler time-frequency diagrams, and generate a sample training set based on the multiple sample micro-Doppler time-frequency diagrams;

[0204] (C.4) Input each sample micro-Doppler time-frequency diagram in the sample training set into the preset classification model in turn to obtain the predicted class label corresponding to each sample data;

[0205] (C.5) Obtain the sample class label of the sample data, and determine the target loss based on the difference between the predicted class label and the sample class label;

[0206] (C.6) Based on the target loss, adjust the parameters of the preset classification model to obtain the classification model.

[0207] Among them, the first echo time-frequency diagram can be an image obtained by performing time-frequency transformation on the real echo signal collected by a sonar device in the actual environment.

[0208] Among them, the diffusion model can be a time-series diffusion model, which is used to generate new samples (second echo time-frequency diagrams) with the same statistical characteristics as the original sample echo signal. In this way, the problem of less sonar echo data can be effectively solved, and the accuracy of underwater target recognition can be improved.

[0209] Among them, the second echo time-frequency diagram can be simulated data generated by the diffusion model, representing the image obtained by performing time-frequency transformation on the echo signal under different micro-motion forms (such as vibration, rotation, etc.), which is used to enhance the data volume and diversity of the training set, so that the preset model has better generalization performance.

[0210] Among them, the actual environment can be a real underwater environment.

[0211] Among them, the sample micro-Doppler component can be the micro-Doppler feature part extracted from the first echo time-frequency diagram and the second echo time-frequency diagram respectively. The sample micro-Doppler component reflects the frequency changes generated by the local parts of the corresponding underwater target (such as engine vibration, rotor blade rotation, etc.).

[0212] Among them, the sample micro-Doppler time-frequency diagram can be a partial time-frequency diagram containing only the micro-Doppler component extracted from the first echo time-frequency diagram and the second echo time-frequency diagram, which shows the distribution of the micro-motion characteristics of the underwater target in time and frequency, removing the main translational frequency components and other noise interferences.

[0213] Among them, the preset classification model can be a deep learning model, such as ResNet-18, etc., for classifying the input sample micro-Doppler time-frequency diagram.

[0214] Among them, the predicted class label can be the class prediction result output by the preset classification model after processing the input sample micro-Doppler time-frequency diagram. Each predicted class label represents the probability that the preset classification model believes that the input data belongs to the corresponding micro-motion category.

[0215] Among them, the sample class label can be the actual class label of each sample data in the sample training set, indicating the true class to which the sample data belongs, and is used to evaluate the performance of the model.

[0216] Among them, the target loss can be a measure of the difference between the predicted class label of the preset classification model and the sample class label. The target loss can be cross-entropy loss, etc., for guiding the adjustment of the parameters of the preset component model to minimize the error.

[0217] Please refer to Figure 6 , for an introduction to the process of training the classification model. Further, after obtaining the sample echo signals of the underwater target, the sample echo signals can be band-pass filtered and preprocessed, and when there is second-order translation in the sample time-frequency sub-image, translational compensation is performed, and then the sample micro-Doppler component is extracted to generate the sample micro-Doppler time-frequency diagram corresponding to each sample echo signal. Further, based on the multiple sample micro-Doppler time-frequency diagrams corresponding to multiple sample echo signals, a test set, a validation set, and a training set can be generated.

[0218] Further, a preset diffusion model can also be used to perform sample data augmentation based on multiple first echo time-frequency diagrams of various micro-motion forms to obtain multiple second echo time-frequency diagrams. After processing the multiple second echo time-frequency diagrams to obtain the corresponding multiple sample micro-Doppler time-frequency diagrams, they can be added to the training set for data augmentation. Then, the training set is used to train the preset classification model to obtain the recognition result output by the preset classification model.

[0219] Exemplarily, multiple first echo time-frequency diagrams corresponding to the actual environment can be divided into a test set, a validation set, and a training set. After generating multiple second echo time-frequency diagrams through a trained diffusion model, the multiple second echo time-frequency diagrams are added to the training set for data augmentation, and a preset classification model is trained based on the training set.

[0220] In some embodiments, the process of extracting the sample micro-Doppler components from the first echo time-frequency diagram and the second echo time-frequency diagram, and generating the sample micro-Doppler time-frequency diagram according to the sample micro-Doppler components is the same as the process of extracting the micro-Doppler components from the echo time-frequency diagram and generating the micro-Doppler time-frequency diagram in the above text. The embodiments of the present application will not elaborate on the processing process of the sample micro-Doppler components in the training process, and the specific process can refer to the foregoing processing process.

[0221] In some embodiments, the target loss can be mean square error loss, cross-entropy loss, squared absolute error loss, etc. The specific form of the loss function can be determined according to the actual situation. Taking the target loss as cross-entropy loss as an example, the target loss L can be calculated by the following formula:

[0222] ;

[0223] where is the number of categories, is the actual sample category label, is the probability of the predicted category label i predicted by the preset classification model.

[0224] Furthermore, the gap between the predicted category label predicted by the preset classification model and the sample category label can be determined based on the target loss, and the parameters of the preset classification model can be adjusted based on the target loss until the preset convergence condition is reached, and then the trained classification model can be obtained. The convergence condition can be that the number of training times reaches a preset number, such as 500 times, 1000 times, etc., or the target loss is less than the loss threshold. The embodiments of the present application do not make specific limitations on this.

[0225] By training the preset classification model with the sample micro-Doppler time-frequency diagrams of the second echo time-frequency diagrams obtained after data augmentation for each category, the recognition ability of the classification model for micro-Doppler features can be effectively enhanced, so as to improve the accuracy and efficiency of recognition, thereby providing important support for fields such as marine science, security monitoring, and resource management.

[0226] In some embodiments, to solve the problem of being unable to effectively classify the micro - motion forms of underwater targets, the present application considers the temporal characteristics of echoes of different basic micro - motion forms and uses a time - series diffusion model to enhance the data of samples, greatly reducing the requirement for the number of original samples and effectively improving the recognition accuracy. Exemplarily, (C.2) may include:

[0227] (C.1.1) For each first echo time - frequency diagram, convert the first echo time - frequency diagram into a one - dimensional time series;

[0228] (C.1.2) Input the one - dimensional time series into a preset diffusion model so that the diffusion model expands based on the first echo time - frequency diagram to obtain multiple second echo time - frequency diagrams with the same micro - motion form as the first echo time - frequency diagram; wherein, the first dimension of each convolutional kernel in the diffusion model is adapted to the dimension of the one - dimensional time series.

[0229] Wherein, the first echo signal may be the original sonar echo data obtained from an underwater target.

[0230] Wherein, the first echo time - frequency diagram may be a two - dimensional image obtained by performing time - frequency transformation (such as short - time Fourier transform or wavelet transform) on the first echo signal. Different first echo signals contain various micro - motion forms (such as vibration, rotation, tumbling, and coning) of the underwater target as well as information such as environmental noise.

[0231] Wherein, the one - dimensional time series may be in the form of converting the first echo time - frequency diagram into a one - dimensional array. By performing projection processing or other processing on the first echo time - frequency diagram along a specific direction, the two - dimensional first echo time - frequency diagram can be compressed into a one - dimensional array.

[0232] Wherein, the first dimension of the convolutional kernel may be the size of the first dimension for processing input data in the convolutional layer of the diffusion model. For one - dimensional time - series input, the first dimension of the convolutional kernel should be adapted to the length of the time series.

[0233] Please refer to Figure 7 . Exemplarily, Figure 7 is the structural diagram of the diffusion model. Specifically, represents the input data of the diffusion model, which may be the original, unprocessed one - dimensional time series. Starting from , during the forward diffusion process of the diffusion model, noise can be gradually added to the data to generate a series of data with noise to gradually change the data distribution from the original distribution to a Gaussian noise distribution.

[0234] Furthermore, at each time step t, a neural network (U-net) can be used to predict the noise added in that step. The U-net learns to remove noise from partially noisy data, thereby generating data closer to the original distribution.

[0235] Furthermore, starting from the noisy data noise is gradually removed, generating a series of data that gradually approaches the original distribution to facilitate the simulation of the process of gradually recovering from the Gaussian noise distribution to the original data distribution. After that, at each time step t, the loss function is calculated to measure the difference between the noise predicted by the model and the actually added noise . By minimizing these losses, the U-net learns to more accurately predict the noise added in each step. Finally, the data generated by the diffusion model should be new data that is similar to or conforms to the one-dimensional time series of the original input . In the above way, the diffusion model can learn to generate high-quality new samples from noisy data, thereby solving the problem of data scarcity and improving the accuracy of underwater target recognition.

[0236] In some embodiments, in order to use the diffusion model for data augmentation and generation, the first echo time-frequency map can be converted into a one-dimensional time series so that the time-frequency data adapts to the diffusion model, reducing the time and resource consumption of the diffusion model inference.

[0237] In some embodiments, the pixel values of the first echo time-frequency map can be arranged in row or column order to form a one-dimensional time series, so that the diffusion model can perform time series analysis and capture the time dependence of the echo signal by simulating the evolution of the data over time.

[0238] In some embodiments, since the diffusion model is usually used to process two-dimensional or three-dimensional image data and its convolutional kernels are designed as two-dimensional or three-dimensional, therefore, in the process of generating samples through the diffusion model in this application, the first dimension (spatial dimension) of the convolutional kernel is no longer applicable. To adapt to the processing of one-dimensional time series, the first dimension of the convolutional kernel is set to 1, which matches the characteristics of the time series data, facilitating the application of convolutional operations by the diffusion model on the time series and capturing the temporal features in the sequence (here it is considered that the sonar echo data has time dependence, that is, the value at the current time point is usually related to the value at the previous or multiple previous time points), and generating a more accurate second echo time-frequency map.

[0239] In some embodiments, in order to enable the classification model to comprehensively capture the complex temporal features and dynamic behaviors in the samples, diffusion can be performed based on the first echo time-frequency diagrams corresponding to various micro-motion forms to obtain a large number of second echo time-frequency diagrams. After training the preset classification model with multiple second echo time-frequency diagrams, the classification model can better adapt to complex environments and improve its performance in real underwater environments.

[0240] By applying a diffusion model to enhance the underwater target recognition dataset based on the first echo time-frequency diagrams of various micro-motion forms and performing temporal processing on the first echo time-frequency diagrams, the diffusion model can focus on temporal feature extraction, thereby generating rich samples, effectively solving the problems of few and difficult-to-obtain underwater samples, improving the recognition performance and accuracy of the model, and optimizing the use of computing resources, facilitating the subsequent improvement of the robustness of the classification model through a large number of samples.

[0241] Please refer to Figure 8 , the embodiment of the present application also provides an underwater target recognition device, which can implement the above-mentioned underwater target recognition method. The underwater target recognition device includes:

[0242] An acquisition module 81, configured to acquire an initial echo signal reflected by an underwater target for a sonar signal and generate an echo time-frequency diagram based on the initial echo signal;

[0243] A first determination module 82, configured to determine the micro-motion period of the underwater target according to the time-frequency curve in the echo time-frequency diagram and determine at least one time-frequency sub-image corresponding to the micro-motion period from the echo time-frequency diagram;

[0244] A construction module 83, configured to project the time-frequency sub-image in multiple directions to obtain multiple projection intensity distributions of the time-frequency sub-image in different projection directions, and construct an energy distribution image based on the multiple projection intensity distributions; wherein, there are overlapping projections between different projection intensity distributions in the energy distribution image;

[0245] A second determination module 84, configured to determine a target energy region with the highest cumulative projection intensity value from the energy distribution image, and obtain a sine curve sequence corresponding to the micro-Doppler component in the time-frequency curve based on the target position of the target energy region in the energy distribution image;

[0246] A generation module 85, configured to extract the micro-Doppler component from the echo time-frequency diagram based on the sine curve sequence and generate a micro-Doppler time-frequency diagram according to the micro-Doppler component;

[0247] An input module 86, configured to input the micro-Doppler time-frequency diagram into a pre-trained classification model to obtain the recognition result of the underwater target.

[0248] The specific implementation manner of the underwater target recognition device is basically the same as the specific embodiment of the above-mentioned underwater target recognition method, and will not be elaborated here. On the premise of meeting the requirements of the embodiments of the present application, other functional modules can also be set in the underwater target recognition device to implement the underwater target recognition method in the above embodiments.

[0249] The embodiments of the present application also provide a computer device, which includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, it implements the above-mentioned underwater target recognition method. The computer device can be any intelligent terminal including a tablet computer, an in-vehicle computer, etc.

[0250] Please refer to Figure 9 , Figure 9 which schematically shows the hardware structure of a computer device in another embodiment. The computer device includes:

[0251] A processor 91, which can be implemented in ways such as a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided by the embodiments of the present application;

[0252] A memory 92, which can be implemented in forms such as a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 92 can store an operating system and other application programs. When implementing the technical solutions provided by the embodiments of this specification through software or firmware, the relevant program codes are stored in the memory 92, and the processor 91 is called to execute the underwater target recognition method of the embodiments of the present application;

[0253] An input / output interface 93, which is used to implement information input and output;

[0254] A communication interface 94, which is used to implement communication interaction between this device and other devices, and can achieve communication through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.);

[0255] A bus 95, which transmits information between various components of the device (such as the processor 91, the memory 92, the input / output interface 93, and the communication interface 94);

[0256] Among them, the processor 91, the memory 92, the input / output interface 93, and the communication interface 94 achieve communication connections with each other inside the device through the bus 95.

[0257] An embodiment of the present application further provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the above-mentioned underwater target recognition method.

[0258] As a non-transitory computer-readable storage medium, the memory can be used to store non-transitory software programs and non-transitory computer-executable programs. In addition, the memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state storage devices. In some embodiments, the memory may optionally include memories remotely provided with respect to the processor, and these remote memories can be connected to the processor through a network. Examples of the above networks include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0259] The embodiments described in the embodiments of the present application are for more clearly illustrating the technical solutions of the embodiments of the present application, and do not constitute a limitation on the technical solutions provided by the embodiments of the present application. Those skilled in the art will know that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of the present application are equally applicable to similar technical problems.

[0260] Those skilled in the art can understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and may include more or fewer steps than those shown in the figures, or combine certain steps, or different steps.

[0261] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0262] Those of ordinary skill in the art can understand that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, and appropriate combinations thereof.

[0263] In the description of the present application and the above-mentioned drawings, terms such as "first", "second", "third", "fourth", etc. (if any) are used to distinguish similar objects and do not necessarily describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances so that the embodiments of the present application described herein can be implemented in an order different from those illustrated or described herein. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that comprises a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or devices.

[0264] It should be understood that in the present application, "at least one (item)" and "several" mean one or more, and "a plurality" means two or more. "And / or" is used to describe the association relationship of associated objects and indicates that three relationships may exist. For example, "A and / or B" may mean: only A exists, only B exists, and both A and B exist at the same time. Among them, A and B can be singular or plural. The character " / " generally means that the associated objects before and after are in an "or" relationship. "At least one (one)" or a similar expression thereof refers to any combination of these items, including any combination of single items (ones) or plural items (ones). For example, at least one (one) of a, b, or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0265] In several embodiments provided in the present application, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative. For example, the above-mentioned division of units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection to each other can be through some interfaces, and the indirect coupling or communication connection of devices or units can be in electrical, mechanical, or other forms.

[0266] The units described above as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they can be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0267] In addition, in each embodiment of the present application, each functional unit may be integrated into a processing unit, may exist separately as individual physical units, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of a software functional unit.

[0268] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it may be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or all or part of this technical solution, may be embodied in the form of a software product. This computer software product is stored in a storage medium and includes multiple instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods in each embodiment of the present application. The aforementioned storage medium includes: various media that can store programs such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs.

[0269] The preferred embodiments of the embodiments of the present application have been described above with reference to the accompanying drawings. However, this does not limit the scope of the rights of the embodiments of the present application. Any modification, equivalent replacement, and improvement made by those skilled in the art without departing from the scope and essence of the embodiments of the present application shall fall within the scope of the rights of the embodiments of the present application.

Claims

1. A method for underwater target recognition, characterized in that: The method comprises: Acquire an initial echo signal reflected by an underwater target in response to a sonar signal, and generate an echo time-frequency diagram based on the initial echo signal; Determine the micro-motion period of the underwater target according to the time-frequency curve in the echo time-frequency diagram, and determine at least one time-frequency sub-image corresponding to the micro-motion period from the echo time-frequency diagram; Projecting the time-frequency sub-image in multiple directions to obtain multiple projection intensity distributions of the time-frequency sub-image in different projection directions, and constructing an energy distribution image based on the multiple projection intensity distributions; Wherein, there are overlapping projections between different projection intensity distributions in the energy distribution image; Determine a target energy region having the highest cumulative projection intensity value from the energy distribution image, and obtain a sinusoidal curve sequence corresponding to a micro-Doppler component in the time-frequency curve based on a target position of the target energy region in the energy distribution image; Extracting a micro-Doppler component from the echo time-frequency diagram based on the sinusoidal curve sequence, and generating a micro-Doppler time-frequency diagram according to the micro-Doppler component; The micro-Doppler time-frequency diagram is input into a pre-trained classification model to obtain a recognition result of the underwater target.

2. The underwater target recognition method according to claim 1, characterized in that: The step of generating an echo time-frequency diagram based on the initial echo signal comprises: Acquiring a radial displacement velocity of the underwater target, a sound velocity of an underwater environment corresponding to the underwater target, and a carrier frequency of the sonar signal, and determining a predicted Doppler frequency of the underwater target based on the radial displacement velocity, the sound velocity, and the carrier frequency; Obtaining a preset filter range coefficient, and setting upper and lower cutoff frequencies of a bandpass filter based on the carrier frequency, the filter range coefficient and the predicted Doppler frequency; The initial echo signal is filtered using the upper and lower cutoff frequencies to obtain a first echo signal; performing normalization processing on the first echo signal to obtain a second echo signal; Perform time-frequency transformation on the second echo signal to obtain an echo time-frequency diagram corresponding to the initial echo signal.

3. The underwater target recognition method according to claim 2, characterized in that: Before the first echo signal is normalized to obtain the second echo signal, the method further includes: Sequentially intercepting signals to be processed from the first echo signal through a preset sliding window; For the signal to be processed corresponding to each sliding window, obtain the mean and standard deviation of the signal to be processed, and set a threshold range of the signal to be processed based on the mean and the standard deviation; When a target signal value in the signal to be processed is outside the threshold range, replacing the target signal value with the mean value; After the processing of the multiple signals to be processed contained in the first echo signal is completed, a processed first echo signal is obtained.

4. The underwater target recognition method according to claim 1, characterized in that: The step of determining the micro-motion period of the underwater target according to the time-frequency curve in the echo time-frequency diagram comprises: Performing autocorrelation analysis on the echo time-frequency diagram to obtain an autocorrelation function corresponding to the initial echo signal; wherein the autocorrelation function is used to measure the degree of overlap of the time-frequency curves under different time delays; Obtaining a first target peak value of the autocorrelation function; Based on the position of the first target peak, the micro-motion period of the underwater target is determined.

5. The underwater target recognition method according to claim 1, characterized in that: Before projecting the time-frequency sub-image in multiple directions to obtain multiple projection intensity distributions of the time-frequency sub-image in different projection directions and constructing an energy distribution image based on the multiple projection intensity distributions, the method further includes: Performing a linear transformation on the time-frequency sub-image to obtain a linear component in the time-frequency sub-image; For each straight line component, obtain the slope and intercept of each straight line component, and determine the second-order translation result of the straight line component based on the slope and the intercept; When the second-order translation result indicates that the linear component has a second-order translation, the frequency axis coordinates in the time-frequency sub-image are offset-adjusted based on the slope and the intercept to obtain a processed time-frequency sub-image.

6. The underwater target recognition method according to claim 1, characterized in that: The classification model is trained in the following way: Acquire multiple first echo time-frequency diagrams corresponding to the underwater target in the actual environment; the first echo time-frequency diagrams correspond to multiple micro-motion forms; By using a preset diffusion model, sample expansion is performed based on a plurality of first echo time-frequency diagrams to obtain a plurality of corresponding second echo time-frequency diagrams; Extracting micro-Doppler components from the first echo time-frequency diagram and the second echo time-frequency diagram respectively to obtain a plurality of sample micro-Doppler time-frequency diagrams, and generating a sample training set based on the plurality of sample micro-Doppler time-frequency diagrams; Inputting each sample micro-Doppler time-frequency graph in the sample training set into a preset classification model in turn to obtain a predicted category label corresponding to each sample micro-Doppler time-frequency graph; Obtaining a sample category label of each sample micro-Doppler time-frequency map, and determining a target loss based on a difference between the predicted category label and the sample category label; Based on the target loss, the parameters of the preset classification model are adjusted to obtain a classification model.

7. The underwater target recognition method according to claim 6, characterized in that: The method of performing sample expansion based on a plurality of first echo time-frequency diagrams through a preset diffusion model to obtain a plurality of corresponding second echo time-frequency diagrams includes: For each first echo time-frequency diagram, converting the first echo time-frequency diagram into a one-dimensional time series; The one-dimensional time series is input into a preset diffusion model so that the diffusion model is expanded based on the first echo time-frequency diagram to obtain a plurality of second echo time-frequency diagrams having the same micro-motion form as the first echo time-frequency diagram; wherein the first dimension of each convolution kernel in the diffusion model is adapted to the dimension of the one-dimensional time series.

8. An underwater target recognition device, characterized in that: The device comprises: An acquisition module, used for acquiring an initial echo signal reflected by an underwater target in response to a sonar signal, and generating an echo time-frequency diagram based on the initial echo signal; A first determination module is used to determine the micro-motion period of the underwater target according to the time-frequency curve in the echo time-frequency diagram, and determine at least one time-frequency sub-image corresponding to the micro-motion period from the echo time-frequency diagram; A construction module, used to project the time-frequency sub-image in multiple directions to obtain multiple projection intensity distributions of the time-frequency sub-image in different projection directions, and construct an energy distribution image based on the multiple projection intensity distributions; wherein there are overlapping projections between different projection intensity distributions in the energy distribution image; A second determination module is used to determine a target energy region with the highest cumulative projection intensity value from the energy distribution image, and obtain a sinusoidal curve sequence corresponding to a micro-Doppler component in the time-frequency curve based on a target position of the target energy region in the energy distribution image; A generating module, configured to extract a micro-Doppler component from the echo time-frequency diagram based on the sinusoidal curve sequence, and generate a micro-Doppler time-frequency diagram according to the micro-Doppler component; An input module is used to input the micro-Doppler time-frequency diagram into a pre-trained classification model to obtain a recognition result of the underwater target.

9. A computer device, characterized in that: The computer device includes a memory and a processor, the memory stores a computer program, and the processor implements the underwater target recognition method according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the underwater target recognition method according to any one of claims 1 to 7 is implemented.

Images