Sonar Image Semantic Communication Method, Apparatus, Device, Storage Medium, and Computer Program Product

Abstract

The present application relates to the field of link communication technologies, and discloses a sonar image semantic communication method, apparatus, device, storage medium, and computer program product, including: obtaining an original sonar image, performing semantic encoding on the original sonar image according to a semantic network, and extracting semantic features; combining the semantic features with the signal-to-noise ratio information of a channel to obtain channel coding output vector data; compressing the channel coding output vector data based on a quantization module to obtain target integer data, and sending the target integer data to a network receiving end through the channel; the network receiving end restores the received target integer data to the channel coding output vector data based on a quantization factor, and performs channel decoding and semantic decoding on the channel coding output vector data to obtain a target sonar image. The sonar image is efficiently transmitted back to a land base station by using semantic communication technology, and challenges brought by scarce sonar image data and limited channel bandwidth for information transmission back are solved.

Description

Technical Field

[0001] The present application relates to the technical field of link communication, and in particular, to a sonar image semantic communication method, apparatus, device, storage medium, and computer program product. Background Art

[0002] When transmitting a large amount of ocean data such as sonar images or videos back to land via a low-earth orbit satellite as a relay for a sea-satellite-land link in the open sea, the bandwidth of the link channel is limited, the channel quality changes dynamically, and the communication quality is poor. Summary of the Invention

[0003] The main purpose of the present application is to provide a sonar image semantic communication method, apparatus, device, storage medium, and computer program product, aiming to solve the technical problem that when transmitting a large amount of ocean data such as sonar images or videos back to land via a low-earth orbit satellite as a relay for a sea-satellite-land link in the open sea, the bandwidth of the link channel is limited, the channel quality changes dynamically, and the communication quality is poor.

[0004] To achieve the above object, the present application proposes a sonar image semantic communication method, which is applied to a network sending end and includes:

[0005] Obtain an original sonar image, and perform semantic encoding on the original sonar image according to a semantic network to extract semantic features;

[0006] Combine the semantic features with the signal-to-noise ratio information of the channel to obtain channel-coded output vector data;

[0007] Compress the channel-coded output vector data based on a quantization module to obtain target integer data, and send the target integer data to a network receiving end through the channel;

[0008] The network receiving end restores the received target integer data to the channel-coded output vector data based on a quantization factor, and performs channel decoding and semantic decoding on the channel-coded output vector data to obtain a target sonar image.

[0009] Optionally, the step of obtaining an original sonar image and performing semantic encoding on the original sonar image according to a semantic network to extract semantic features includes:

[0010] Obtain an original sonar image collected by an ocean operation device;

[0011] Divide the original sonar image into blocks according to a preset network hyperparameter setting table in the semantic network to obtain small-sized sonar image blocks;

[0012] Transport the sonar image block into a multi-layer Swin Transformer module to extract the semantic features of the original sonar image.

[0013] Optionally, the step of transporting the sonar image block into a multi-layer Swin Transformer module to extract the semantic features of the original sonar image includes:

[0014] Use the segmented sonar image blocks as input and transport them into a multi-layer Swin Transformer module;

[0015] In the multi-layer Swin Transformer module, alternately use the multi-head attention mechanism with a fixed window and the multi-head attention mechanism with a moving window to capture the initial semantic features of the sonar image block locally and globally;

[0016] Through residual connection and layer normalization, calculate and adjust the initial semantic features, and integrate the adjusted initial semantic features to obtain the semantic features of the original sonar image.

[0017] Optionally, before the step of obtaining the original sonar image collected by the ocean operation equipment, it further includes:

[0018] Select a synthetic aperture radar dataset for pre-training on the semantic network to obtain pre-trained network parameters;

[0019] Based on transfer learning and the fine-tuning method, use a preset sonar image dataset to adjust the pre-trained network parameters to obtain target network parameters;

[0020] Apply the target network parameters to the semantic network to construct the network hyperparameter setting table and the Swin Transformer module.

[0021] Optionally, the step of combining the semantic features with the signal-to-noise ratio information of the channel to obtain the channel coding output vector data includes:

[0022] Collect the signal-to-noise ratio information of the current channel;

[0023] Based on the global average pooling function, obtain the global statistical information of the semantic features on the entire feature map, and splice the global statistical information with the signal-to-noise ratio information to obtain context information;

[0024] Transport the context information into a fully connected neural network, and calculate a set of weight factors based on the semantic features according to the signal-to-noise ratio attention module;

[0025] Multiply the semantic features by the corresponding weight factors respectively to obtain the channel coding output vector data.

[0026] In addition, to achieve the above object, the present application also proposes a sonar image semantic communication method, which is applied to a network receiving end and includes:

[0027] Receiving target integer data sent by a network sending end, where the network sending end is used to perform semantic encoding on an original sonar image, extract semantic features, and after combining the semantic features with the signal-to-noise ratio information of a channel, compress them into the target integer data;

[0028] Restoring the target integer data to channel coding output vector data based on a quantization factor;

[0029] Performing channel decoding on the channel coding output vector data according to the signal-to-noise ratio information, and reconstructing the semantic features of the sonar image;

[0030] Performing semantic decoding on the semantic features of the sonar image, and reconstructing to obtain a target sonar image.

[0031] In addition, to achieve the above object, the present application also proposes a sonar image semantic communication device, which includes:

[0032] A feature extraction module, configured to obtain an original sonar image and perform semantic encoding on the original sonar image according to a semantic network to extract semantic features;

[0033] A data generation module, configured to combine the semantic features with the signal-to-noise ratio information of a channel to obtain channel coding output vector data;

[0034] A data transmission module, configured to compress the channel coding output vector data based on a quantization module to obtain target integer data, and send the target integer data to a network receiving end through a channel;

[0035] An image reconstruction module, configured to restore the received integer data to the channel coding output vector data based on a quantization factor, and perform channel decoding and semantic decoding on the channel coding output vector data to obtain a target sonar image.

[0036] In addition, to achieve the above object, the present application also proposes a sonar image semantic communication device, which includes: a memory, a processor, and a computer program stored on the memory and executable on the processor, where the computer program is configured to implement the steps of the sonar image semantic communication method as described above.

[0037] In addition, to achieve the above object, the present application also proposes a storage medium, which is a computer-readable storage medium. A computer program is stored on the storage medium, and when the computer program is executed by a processor, the steps of the sonar image semantic communication method described above are implemented.

[0038] In addition, to achieve the above object, the present application also provides a computer program product, which includes a computer program. When the computer program is executed by a processor, the steps of the sonar image semantic communication method described above are implemented.

[0039] In the present application, an original sonar image is obtained, and semantic encoding is performed on the original sonar image according to a semantic network to extract semantic features; the semantic features are combined with the signal-to-noise ratio information of the channel to obtain channel coding output vector data; the channel coding output vector data is compressed based on a quantization module to obtain target integer data, and the target integer data is sent to a network receiving end through the channel; the network receiving end restores the received target integer data to the channel coding output vector data based on a quantization factor, and performs channel decoding and semantic decoding on the channel coding output vector data to obtain a target sonar image. By constructing a semantic encoding and decoding network, more efficient detailed capture of semantics is performed on the original signal, and the sonar image is efficiently transmitted back to the land base station using semantic communication technology, and the integrity of semantics is more accurately maintained during the transmission process, thereby improving the transmission efficiency of communication. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] The accompanying drawings herein are incorporated into the specification and form a part of the specification, showing embodiments consistent with the present application, and are used together with the specification to explain the principles of the present application.

[0041] To more clearly illustrate the technical solutions in the embodiments of the present application or in the prior art, the following will briefly introduce the accompanying drawings required for describing the embodiments or the prior art. Obviously, for those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0042] Figure 1 It is a schematic flowchart of the first embodiment of the sonar image semantic communication method of the present application;

[0043] Figure 2 It is a structural diagram of the SNR attention module of the present application;

[0044] Figure 3 It is a sea-satellite-land link diagram for sonar image data transmission of the present application;

[0045] Figure 4It is a schematic flowchart of the second embodiment of the sonar image semantic communication method of this application;

[0046] Figure 5 It is a structural diagram of the semantic encoding network of this application;

[0047] Figure 6 It is an internal structural diagram of the Swin Transformer module of this application;

[0048] Figure 7 It is a schematic diagram of the SAR image target of this application;

[0049] Figure 8 It is a schematic diagram of the forward-looking sonar image target of this application;

[0050] Figure 9 It is a schematic flowchart of the third embodiment of the sonar image semantic communication method of this application;

[0051] Figure 10 It is a flowchart of the sonar image semantic communication of this application;

[0052] Figure 11 It is a structural diagram of the semantic decoding network of this application;

[0053] Figure 12 It is a comparison chart of the pre-training effect of this application;

[0054] Figure 13 It is a comparison chart of the MS-SSIM experimental results;

[0055] Figure 14 It is a comparison chart of the PSNR experimental results;

[0056] Figure 15 It is a schematic diagram of the module structure of the sonar image semantic communication device in the embodiment of this application;

[0057] Figure 16 It is a schematic diagram of the device structure of the hardware operating environment involved in the sonar image semantic communication method in the embodiment of this application.

[0058] The realization of the purpose, functional features and advantages of this application will be further described with reference to the embodiments and the accompanying drawings. Specific embodiments

[0059] It should be understood that the specific embodiments described herein are only used to explain the technical solutions of this application and are not used to limit this application.

[0060] For a better understanding of the technical solutions of this application, the following will be described in detail with reference to the accompanying drawings of the specification and specific embodiments.

[0061] The main solution of the embodiment of the present application is: obtaining an original sonar image, performing semantic encoding on the original sonar image according to a semantic network, and extracting semantic features; combining the semantic features with the signal-to-noise ratio information of a channel to obtain channel-coded output vector data; compressing the channel-coded output vector data based on a quantization module to obtain target integer data, and sending the target integer data to a network receiving end through the channel; the network receiving end restores the received target integer data to the channel-coded output vector data based on a quantization factor, and performs channel decoding and semantic decoding on the channel-coded output vector data to obtain a target sonar image.

[0062] Since the data transmission from the open sea can usually be achieved through sea surface nodes and low Earth orbit (LEO) satellites as media, currently, the data transmission service between ships and shore bases on the sea mainly relies on traditional maritime communication systems, such as the Global Maritime Distress and Safety System, Navtex, and the Automatic Identification System. However, these systems can only provide low data rate communication. In addition, the harsh marine propagation environment and limited communication coverage have caused serious obstacles to the communication from the sea surface to the land. LEO can be used as a relay to achieve the backhaul of open sea ocean data for the sea surface-satellite-land link due to its wide enough coverage. However, because the channel bandwidth of this link is limited and the channel quality changes dynamically, the communication quality is poor, especially for the rapid backhaul of a large amount of marine data such as sonar images or videos. In addition, semantic communication is a potential solution to the channel bandwidth limitation. Nevertheless, most of the existing semantic communication methods based on deep learning are trained for a fixed channel environment, which leads to a rapid decline in communication performance when the channel conditions change. In addition, deep learning requires a large amount of training data to make the model have generalization. However, due to factors such as the high cost of collecting marine sonar images, the sonar image data set is very limited, and the existing semantic communication methods based on deep learning are difficult to be applied to the transmission of sonar images.

[0063] The present application provides a sonar image semantic communication method, which is oriented to the communication link of the sea surface-satellite-land, and uses semantic communication technology to efficiently backhaul sonar images to the land-based station. It solves the challenges brought by the scarcity of sonar image data and the limited channel bandwidth for information backhaul.

[0064] Based on this, the embodiment of the present application provides a sonar image semantic communication method, referring to Figure 1 , Figure 1 is the schematic flowchart of the first embodiment of the sonar image semantic communication method of the present application.

[0065] It should be noted that the execution entity of this embodiment can be a computing service device with data processing, network communication, and program running functions, or an electronic device capable of implementing the above functions, etc. Hereinafter, taking the network sending end as an example, this embodiment and the following embodiments will be described.

[0066] In this embodiment, the sonar image semantic communication method is applied to the network sending end, including:

[0067] Step S10, obtain the original sonar image, and perform semantic encoding on the original sonar image according to the semantic network to extract semantic features.

[0068] It should be noted that the original sonar image is an acoustic wave reflection image in water or other media obtained by a sonar device, which contains a large amount of original data and information, and is usually presented in the form of an image, used to represent underwater terrain, object distribution, etc. The semantic network is a network structure for extracting the semantic information of the original sonar image. Semantic encoding is the process of converting the original data or information into a code or representation form with specific semantic meanings, which involves converting features such as pixel values, textures, and shapes of the image into feature vectors or labels with specific semantic meanings. Semantic features refer to the features with specific semantic meanings in the image, including the undulation of underwater terrain, the shape and size of objects, the intensity of reflected acoustic waves, etc.

[0069] It can be understood that the semantic communication method performs more efficient compression by extracting semantics from the original signal and more accurately maintains the integrity of semantics during the transmission process, thereby improving the transmission efficiency of communication. Compared with the traditional digital communication method, this method is more suitable for communication scenarios with limited bandwidth such as sea-satellite-land.

[0070] Step S20, combine the semantic features with the signal-to-noise ratio information of the channel to obtain the channel coding output vector data.

[0071] It should be noted that the communication channel quality of sea-satellite-ground often changes dynamically. In addition to the path loss caused by long-distance propagation, the absorption loss of the atmosphere and weather changes also affect the signal quality. The uncertainties of these external environments require that the communication system should have the ability to adaptively adjust according to the channel conditions. By combining the semantic features with the signal-to-noise ratio (SNR) information of the channel, resource allocation can be adaptively performed according to the differences in channel conditions to achieve optimal data transmission under limited bandwidth.

[0072] Of course, in order to make the scheme more robust in a dynamic channel environment. The step S20 may include:

[0073] Collect the signal-to-noise ratio information of the current channel; obtain the global statistical information of the semantic features on the entire feature map based on the global average pooling function, and splice the global statistical information with the signal-to-noise ratio information to obtain context information; send the context information into a fully connected neural network, and calculate a set of weight factors based on the semantic features according to the signal-to-noise ratio attention module; multiply the semantic features by the corresponding weight factors respectively to obtain the channel coding output vector data.

[0074] It should be noted that global average pooling is a special pooling operation that averages the entire feature map to obtain the global statistical information of each channel, which can reduce the dimension of the feature map while retaining important global information. The weight factor is a coefficient used in the neural network to adjust the importance between different features or neurons.

[0075] It should be understood that in the case of poor channel quality (low SNR), it is necessary to amplify some key features through the weight factor to ensure that they will not be overwhelmed by noise during transmission. On the contrary, in the case of good channel quality (high SNR), the intensity of the features can be appropriately reduced to reduce unnecessary data transmission. The weight factor enables the model to differentially process different semantic features, improve the transmission quality of important features, and at the same time may reduce the priority of less critical features.

[0076] It can be understood that by using global average pooling and a fully connected network, the computational overhead can be effectively reduced and fast coding adjustment can be achieved. And the generated weighting factor makes the model more robust in a dynamic channel environment. For the sake of easy understanding, taking the actual situation as an example, but not limiting the present application, in an example, refer to Figure 2 , Figure 2 is the structure diagram of the SNR attention module of the present application. To address the challenge of dynamically changing channel quality, a method that can adaptively adjust channel coding according to the signal-to-noise ratio (SNR) is proposed. During the semantic model optimization process, the SNR value range of the simulation channel is set to [1, 13] dB, covering a wide range of scenarios from low SNR to high SNR. By training the neural network across the entire SNR range, the adaptability of the model under different SNR conditions has been significantly improved. The SNR attention module is proposed based on the principle of the attention mechanism, where FC is the fully connected layer, and Relu and Sigmoid are activation functions. The SNR attention mechanism can weight the input features, thereby enhancing the model's ability to capture key information. At the same time, it receives the feature output from the semantic coding and the SNR information of the channel as inputs. Based on these inputs, the SNR attention module calculates a set of weight factors through the neural network, and these factors assign different importance to different features according to the current channel conditions.

[0077] Let represent the semantic coding features, where is the number of channels of the features, is the size of the features. Similarly, let represent the weighted features generated by the SNR attention module. To effectively utilize the input features and adjust according to the signal-to-noise ratio, the SNR attention module first uses the global average pooling function, which can capture the global statistical information of the entire feature map, and then concatenates with the signal-to-noise ratio to form context information. These context information are fed into a fully connected neural network to generate a weight factor. Finally, by multiplying the feature with the weight factor, the feature after channel coding is obtained.

[0078] Step S30, compress the channel coding output vector data based on the quantization module to obtain the target integer data, and send the target integer data to the network receiving end through the channel.

[0079] It should be noted that the sea-surface - satellite - ground channel bandwidth is limited, which requires reducing the size of the data during transmission. During the semantic coding process of sonar images, although the features have been mapped to a compact semantic expression form through dimensionality reduction, their data still exists in the form of 32-bit floating-point numbers, which occupies a large storage space. The purpose of the quantization module is to perform quantization rounding on the low-dimensional semantic vectors while maintaining the key semantic information, and compress the semantic vectors into a smaller bitstream to adapt to the bandwidth limitation.

[0080] In one example, the data type of the low-dimensional semantic representation is converted from 32-bit floating-point type to 8-bit integer type through quantization, reducing the number of transmission bits to one-fourth. The quantization function used is:

[0081]

[0082] where, is the low-dimensional semantics output by the semantic encoder, is the quantization result, is the scale parameter to be learned, is the rounding function. And before semantic decoding at the receiving end, the received semantics need to be dequantized and restored to floating-point data. The dequantization function is:

[0083]

[0084] where, is the vector after being restored to floating-point data, is the bias term to eliminate part of the additive noise.

[0085] Step S40: The network receiving end restores the received target integer data to the channel-coded output vector data based on the quantization factor, and performs channel decoding and semantic decoding on the channel-coded output vector data to obtain the target sonar image.

[0086] It should be noted that the target sonar image is not exactly the same as the original sonar image. Before decoding, it is usually necessary to preprocess the received signal, such as filtering, amplification, etc., to improve the signal quality and detectability. Through the decoding algorithm, the original information vector after channel coding can be restored. These information vectors may still contain some redundant information, but most of the noise and interference have been removed.

[0087] In one example, refer to Figure 3 , Figure 3 is the sea-satellite-land link diagram for sonar image data transmission in this application. When constructing the sea-satellite-ground communication link, LEO is selected as the data transmission intermediary for sonar images. The process is described as follows: 1. After obtaining the underwater sonar image data, ocean operation equipment (such as ships, etc.) uses semantic communication technology to perform semantic coding on the sonar image and uploads the data to the low-earth orbit satellite through a wireless channel. 2. The low-earth orbit satellite forwards the received data to the ground base station using its wide communication coverage. 3. The ground base station uses semantic communication technology to perform semantic decoding on the received data to reconstruct the sonar image data.

[0088] In this embodiment, the original sonar image is obtained, and semantic coding is performed on the original sonar image according to the semantic network to extract semantic features; the semantic features are combined with the signal-to-noise ratio information of the channel to obtain the channel-coded output vector data; the channel-coded output vector data is compressed based on the quantization module to obtain the target integer data, and the target integer data is sent to the network receiving end through the channel; the network receiving end restores the received target integer data to the channel-coded output vector data based on the quantization factor, and performs channel decoding and semantic decoding on the channel-coded output vector data to obtain the target sonar image. By constructing a semantic encoding and decoding network, more efficient detailed capture of semantics is performed on the original signal, and the sonar image is efficiently transmitted back to the land base station using semantic communication technology, and the semantic integrity is more accurately maintained during the transmission process, thereby improving the transmission efficiency of communication.

[0089] Refer to Figure 4 , Figure 4 is the schematic flowchart of the second embodiment of the sonar image semantic communication method in this application. Based on the above first embodiment, the second embodiment of the sonar image semantic communication method in this application is proposed.

[0090] In the second embodiment, the step S10 includes:

[0091] Step S101: Obtain the original sonar image collected by the ocean operation equipment.

[0092] It should be noted that the ocean operation equipment refers to the equipment used for various operations in the ocean environment, including sonar systems, underwater robots, and ocean survey ships, etc.

[0093] Step S102: Divide the original sonar image into blocks according to the preset network hyperparameter setting table in the semantic network to obtain small-sized sonar image blocks.

[0094] It should be noted that dividing the original sonar image into blocks to obtain small-sized sonar image blocks can better capture the imaging relationship between the target object and the acoustic shadow, thereby improving the model's semantic understanding ability of the sonar image.

[0095] It should be understood that the current applications of sonar image semantic communication are relatively limited. Most of the existing image semantic communication network structures are based on convolutional neural networks or Vision Transformer (ViT, Visual Transformer) to extract semantic information. These network structures are mainly designed for optical images and have the following limitations when applied to sonar images: The main targets in sonar images are usually small, and there is a lack of design for small-scale target features; The uniqueness of sonar images lies in their acoustic shadow features, which requires the semantic extraction network to be able to capture both global information and focus on local details.

[0096] In one example, based on the imaging characteristics of sonar images, the hyperparameter settings of the Swin-Transformer network are shown in Table 1.

[0097] Table 1 Network Hyperparameter Setting Table

[0098]

[0099] Among them, the Patch (patch) size is set to 2×2 to meet the semantic feature extraction requirements of small-sized targets in sonar images; the window size is 8×8, which has significant advantages in processing small-sized targets and detail features in sonar images and can extract target features more accurately.

[0100] Step S103: Feed the sonar image blocks into a multi-layer Swin Transformer module to extract the semantic features of the original sonar image.

[0101] It can be understood that the preprocessed and segmented sonar image patches are used as inputs. Utilizing the structure of Swin Transformer, semantic features of the images are extracted through multiple levels of Transformer modules. Each module contains a self-attention mechanism and a multi-layer perceptron (MLP) to capture features at different scales, enhancing the understanding of global information and the ability to capture details in sonar images.

[0102] For ease of understanding, taking the actual situation as an example (but not limiting the present application), in one example, refer to Figure 5 , Figure 5 This is the structural diagram of the semantic encoding network of the present application. First, the source image is segmented into small patches and embedded into the feature space through the Patch Embedding stage. Then, these image patches are sequentially processed through multiple Swin Transformer modules, and each module is followed by a downsampling step to gradually extract and fuse the multi-scale semantic features of the image. Specifically, the image is first processed by a Swin Transformer module 1, and then downsampled; then it passes through Swin Transformer module 2 and is downsampled again; this process is repeated three times in Swin Transformer module 3, and each module is followed by a downsampling step; finally, the image is processed by Swin Transformer module 4 to complete the extraction of semantic features and finally output the semantic representation of the image. This process realizes the effective encoding of the deep semantic information of sonar images through a hierarchical structure and step-by-step downsampling.

[0103] Furthermore, it significantly enhances the understanding of global information and the ability to capture details in sonar images. The step S103 may include:

[0104] Feeding the segmented sonar image patches as inputs into multiple Swin Transformer modules; in the multiple Swin Transformer modules, alternately using the multi-head attention mechanism with a fixed window and the multi-head attention mechanism with a moving window to capture the initial semantic features of the sonar image patches locally and globally; through residual connection and layer normalization, calculating and adjusting the initial semantic features, and integrating the adjusted initial semantic features to obtain the semantic features of the original sonar image.

[0105] It should be understood that in the Swin Transformer module, the window-based multi-head self-attention mechanism (W-MSA) and the shifted window multi-head self-attention mechanism (SW-MSA) are alternately used. The window-based multi-head self-attention mechanism allows the model to capture detailed features within a local area, while the shifted window multi-head self-attention mechanism helps the model cross local windows to capture more extensive context information.

[0106] For ease of understanding, taking the actual situation as an example, but not limiting the present application, in one example, refer to Figure 6 , Figure 6 which is the internal structure diagram of the Swin Transformer module of the present application. In the figure, the window-based multi-head attention module (W-MSA) and the shifted window multi-head attention module (SW-MSA) are alternately used within one module. The "Swin Transformer module" for the input , after the following calculations, obtains :

[0107]

[0108]

[0109]

[0110]

[0111] Among them, MLP (Multilayer Perceptron) is a multi-layer perceptron structure, and LN (Layer Normalization) is a layer normalization operation. In W-MSA, the image patches (Tokens) first go through layer normalization (LN), then are processed by the window-based multi-head attention mechanism, and then go through LN and residual connection again to generate the output . In SW-MSA, the image patches also first go through LN, but then are processed by the shifted window multi-head attention mechanism, which increases the receptive field by moving the window at different positions, and then goes through LN and residual connection again to generate the output . These two mechanisms are alternately used to capture local and cross-window dependencies, so as to extract multi-scale features of the image.

[0112] Of course, in order to solve the problem of underfitting of the feature extraction network caused by the scarcity of sonar image datasets, before the step S101, the following is also included:

[0113] Select a synthetic aperture radar dataset for pre-training on the semantic network to obtain pre-trained network parameters; based on transfer learning, select the fine-tuning method, and use a preset sonar image dataset to adjust the pre-trained network parameters to obtain target network parameters; apply the target network parameters to the semantic network to construct the network hyperparameter setting table and the Swin Transformer module.

[0114] It should be understood that transfer learning is an effective method to solve the problem of insufficient training data in deep learning. It uses rich data in other related fields to pre-train the model parameters of the neural network, and then transfers the trained model to a new model to optimize the learning efficiency of the model. Selecting SAR (Synthetic Aperture Radar) images as pre-training data, the main reasons are as follows: Both sonar images and SAR images are grayscale images, and they show certain similarities in shallow features, such as the edges of targets and the distribution of speckle noise. The targets in SAR images and sonar images are mainly small targets.

[0115] In one example, refer to Figure 7 and Figure 8 , Figure 7 is a schematic diagram of the SAR image target of this application, Figure 8 is a schematic diagram of the forward-looking sonar image target of this application. In the figure, the features of the two images are similar. When using transfer learning to transfer the knowledge pre-trained from SAR data to sonar image processing tasks, assuming the domain of the image distribution is , the task can be expressed as , where is the label space, is the target prediction function, then transfer learning can be defined as: Given a learning task on the domain , this task can obtain relevant learning experiences from a learning task on another domain . The purpose of transfer learning is to obtain knowledge from and and transfer it to to improve the performance of the prediction function . This shows that when the distribution differences between the two image domains are large, the experiences that can be learned from are relatively limited, and the performance improvement of using transfer learning is not obvious.

[0116] In this solution, the SAR dataset SARDet-100K is selected for pre-training on the proposed semantic network. Then, through transfer learning, the pre-trained network parameters are applied to sonar images, and a semantic communication neural network with good generalization performance can be obtained by training with a small number of sonar images. The information of the selected sonar image dataset and SAR image dataset is shown in Table 2. The transfer learning method selected is fine-tuning. Based on the pre-trained model parameters, the model parameters are fine-tuned using the sonar image dataset. Table 2 Dataset Information

[0117]

[0118] As can be seen from the table, both datasets have airplanes and ships in their target categories, and the identity of high-level semantics also reduces the semantic differences between the pre-training datasets. Therefore, selecting this dataset for pre-training the feature network helps the network learn the shallow and deep semantic features of sonar images.

[0119] In this embodiment, the original sonar images collected by marine operation equipment are obtained; the original sonar images are segmented according to the preset network hyperparameter setting table in the semantic network to obtain small-sized sonar image blocks; the sonar image blocks are fed into a multi-layer Swin Transformer module to extract the semantic features of the original sonar images. By introducing a sliding window mechanism and using a smaller Patch size to capture the relationship between the target object and the acoustic shadow, semantic features are extracted from different scales, significantly improving the ability to understand the global information and capture details of sonar images.

[0120] Refer to Figure 9 , Figure 9 which is the flowchart of the third embodiment of the sonar image semantic communication method of this application. Based on the above second embodiment, the third embodiment of the sonar image semantic communication method of this application is proposed.

[0121] In this embodiment, the sonar image semantic communication method is applied to the network receiving end, including:

[0122] Step S40, receiving the target integer data sent by the network sending end, where the network sending end is used to perform semantic encoding on the original sonar images, extract semantic features, and compress them into target integer data after combining the semantic features with the signal-to-noise ratio information of the channel.

[0123] It should be noted that the execution subject of this embodiment can be a computing service device with data processing, network communication, and program running functions or an electronic device capable of implementing the above functions. Hereinafter, the network receiving end is taken as an example to illustrate this embodiment and the following embodiments.

[0124] In a specific example, refer toFigure 10 , Figure 10 This is the flowchart of sonar image semantic communication for this application. First, the sonar image is input as the information source. After the semantic encoding stage, the image information is converted into semantic features. Then, these semantic features are channel-encoded and combined with the channel state information (SNR) for quantization processing to adapt to channel transmission. In the channel, these quantized data are subject to noise interference. At the receiving end, after dequantization, the semantic information of the image is reconstructed through channel decoding and semantic decoding. Finally, through the semantic decoder, the semantic information is converted back into the image form to obtain the reconstructed sonar image. During the whole process, SAR images are used for pre-training to improve the model's learning ability and generalization of sonar image features. This method realizes the effective extraction and transmission of the semantic information of sonar images in the case of scarce sonar image data by utilizing the similarity of SAR images and sonar images in shallow features and their common feature of mainly small targets.

[0125] Step S50: Restore the target integer data to the channel-coded output vector data based on the quantization factor.

[0126] It can be understood that before semantic decoding at the receiving end, the received semantics need to be dequantized, and the compressed data is restored using the quantization factor. , is the scale parameter to be learned, is the vector after being restored to floating-point data, is the bias term to eliminate part of the additive noise.

[0127] Step S60: Perform channel decoding on the channel-coded output vector data according to the SNR information to reconstruct the semantic features of the sonar image.

[0128] It can be understood that after dequantization, the semantic decoding network is used to further process the features to reconstruct the semantic features of the sonar image. The semantic decoding network is usually similar in structure to the semantic encoding network but in the opposite direction, and it can recover the high-level semantic information of the image from the compressed semantic features.

[0129] Step S70: Perform semantic decoding on the semantic features of the sonar image to reconstruct the target sonar image.

[0130] It should be understood that the step of performing semantic decoding on the semantic features of the sonar image to reconstruct the target sonar image is exactly the opposite of the encoding step, and upsampling processing is performed after each module.

[0131] In one example, refer to Figure 11 , Figure 11This is the structural diagram of the semantic decoding network of this application. First, the model receives the semantic information transmitted through the channel. The received semantic information is first processed by the Swin Transformer module 4, and then successively processed by upsampling and the Swin Transformer module 3 (repeated three times), further upsampling and the Swin Transformer module 2, upsampling and the Swin Transformer module 1, and finally the original sonar image is reconstructed through the last upsampling step.

[0132] In a specific scenario, referring to Figure 12 、 Figure 13 and Figure 14 , Figure 12 This is the comparison diagram of the pre-training effect of this application. Figure 13 and Figure 14 are respectively the comparison diagram of the MS-SSIM experimental results and the comparison diagram of the PSNR experimental results. The performance of the present invention is verified with an AWGN (Additive White Gaussian Noise) simulation channel. During the optimization process of the semantic model, the SNR value range of the simulation channel is set to [1, 13] dB, covering a wide range of conditions from low signal-to-noise ratio to high signal-to-noise ratio. The evaluation metrics for verifying the proposed semantic communication method in this scheme are the Peak Signal-to-Noise Ratio (PSNR) and the Multi Scale Structural Similarity Index Measure (MS-SSIM). PSNR calculates the pixel value error between the semantic recovery image and the original image. The larger the value of PSNR, the more similar the two images are, and it is defined as:

[0133]

[0134] where m and n are the height and width of the image, is the pixel value of the original image, is the pixel value of the recovery image, is the maximum pixel value of the image. If the image pixel value is represented by 8-bit binary, then = 255.

[0135] MS-SSIM is an index for measuring the similarity between two images. It compares the brightness, contrast, and structure of the two images. The closer the value of MS-SSIM is to 1, the more similar the two images are. It is defined as:

[0136]

[0137] where, is the brightness measurement, is the contrast measure, is the structure measure. are the contrast measure and structure measure obtained after the image undergoes j - 1 times of downsampling, with a total of M - 1 times of downsampling.

[0138] First, the effectiveness of using SAR images as pre - training data for sonar image tasks is verified. In the experiment, the official training set of the optical image dataset COCO2014 is selected for comparison. It contains 82,783 images and a total of 80 categories. The experimental results are as Figure 12 shown. In the figure, the horizontal axis represents the SNR value of the simulation channel during testing, and the vertical axis represents the PSNR result. It can be seen from the figure that when testing on the sonar image dataset, the network model pre - trained with the SAR dataset achieves better PSNR results at different channel SNRs.

[0139] To verify the effectiveness of the proposed sonar image semantic communication method, the proposed method is compared with the Deep - Joint Source - Channel Coding (Deep - JSCC) scheme based on SwinTransformer and the traditional separate source - channel coding scheme BPG+LDPC (Better Portable Graphics+Low - Density Parity - Check, BPG image compression technology and LDPC error - correcting code). The experimental results are as Figure 13 and Figure 14 shown. The results show that the proposed method is superior to the traditional digital communication method BPG+LDPC under various signal - to - noise ratio (SNR) conditions. Especially under low SNR conditions, the latter can no longer effectively recover the image at the receiver. The Deep - JSCC method considers fixed SNR conditions during optimization, but has poor stability under different SNR conditions and can only obtain better communication performance when the channel conditions are similar to those during model training. In contrast, the method proposed in this scheme shows higher PSNR and MS - SSIM results under different SNR conditions, proving its better adaptability to dynamic changes in channel quality.

[0140] The compression ratio is defined as the ratio of the number of transmitted bytes to the number of source bytes to measure the data transmission efficiency. The smaller the compression ratio, the fewer bytes are required to send the same source. According to the experimental data in Table 3, the sonar image communication method significantly reduces the required number of transmitted bits while maintaining the communication quality, showing higher transmission efficiency compared to traditional communication methods. In addition, after adding the quantization module, this method further reduces the number of transmitted bytes while still maintaining excellent transmission quality. This result indicates that the method is suitable for communication environments with limited channel bandwidth and can achieve efficient data transmission with limited bandwidth resources.

[0141] Table 3 Comparison of PSNR performance of different compression ratio methods

[0142]

[0143] In this embodiment, the target integer data sent by the receiving network sending end is received. The network sending end is used to perform semantic encoding on the original sonar image, extract semantic features, and after combining the semantic features with the signal-to-noise ratio information of the channel, compress them into target integer data, and restore the target integer data to channel coding output vector data based on the quantization factor; perform channel decoding on the channel coding output vector data according to the signal-to-noise ratio information, and reconstruct the semantic features of the sonar image; perform semantic decoding on the semantic features of the sonar image to reconstruct the target sonar image. By constructing a semantic encoding and decoding network, decoding the received target data, restoring the semantic features, and reconstructing the image, fast feedback of a large amount of marine data such as sonar images or videos is achieved, ensuring the accuracy of image transmission.

[0144] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the sonar image semantic communication method of this application. Based on this technical concept, more forms of simple transformations are within the protection scope of this application.

[0145] This application also provides a sonar image semantic communication device. Please refer to Figure 15 , the sonar image semantic communication device includes:

[0146] Feature extraction module 10, which is used to obtain the original sonar image and perform semantic encoding on the original sonar image according to the semantic network to extract semantic features;

[0147] Data generation module 20, which is used to combine the semantic features with the signal-to-noise ratio information of the channel to obtain channel coding output vector data;

[0148] Data transmission module 30, which is used to compress the channel coding output vector data based on the quantization module to obtain target integer data, and send the target integer data to the network receiving end through the channel;

[0149] An image reconstruction module 40, configured to restore the received integer data to the channel-coded output vector data based on a quantization factor, and perform channel decoding and semantic decoding on the channel-coded output vector data to obtain a target sonar image.

[0150] The sonar image semantic communication device provided in this application adopts the sonar image semantic communication method in the above embodiment, and can solve the technical problems of limited bandwidth of the link channel, dynamic change of channel quality, and poor communication quality when transmitting far-sea marine data of a sea-surface-satellite-land link with a low-earth orbit satellite as a relay for a large amount of marine data such as sonar images or videos. Compared with the prior art, the beneficial effects of the sonar image semantic communication device provided in this application are the same as those of the sonar image semantic communication method provided in the above embodiment, and other technical features in the sonar image semantic communication device are the same as the features disclosed in the method of the above embodiment, which will not be elaborated here.

[0151] This application provides a sonar image semantic communication device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the sonar image semantic communication method in Embodiment 1 above.

[0152] Next, refer to Figure 16 , which shows a schematic structural diagram of a sonar image semantic communication device suitable for implementing the embodiments of this application. The sonar image semantic communication device in the embodiments of this application may include, but is not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Descriptions: tablet computers), PMPs (Portable Media Players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 16 The sonar image semantic communication device shown is only an example and should not impose any limitations on the functions and usage scopes of the embodiments of this application.

[0153] As Figure 16As shown, the sonar image semantic communication device may include a processing device 1001 (such as a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM: Read Only Memory) 1002 or the program loaded from the storage device 1003 into the random access memory (RAM: Random Access Memory) 1004. In the RAM 1004, various programs and data required for the operation of the sonar image semantic communication device are also stored. The processing device 1001, the ROM 1002, and the RAM 1004 are connected to each other through a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Generally, the following systems can be connected to the I / O interface 1006: an input device 1007 including, for example, a touch screen, a touch pad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output device 1008 including, for example, a liquid crystal display (LCD: Liquid Crystal Display), a speaker, a vibrator, etc.; a storage device 1003 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 1009. The communication device 1009 can allow the sonar image semantic communication device to communicate with other devices wirelessly or wiredly to exchange data. Although the figure shows a sonar image semantic communication device having various systems, it should be understood that it is not required to implement or have all the shown systems. More or fewer systems can be alternatively implemented or had.

[0154] In particular, according to the embodiments disclosed in the present application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, the embodiments disclosed in the present application include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program contains program codes for performing the methods shown in the flowcharts. In such an embodiment, the computer program can be downloaded and installed from the network through the communication device, or installed from the storage device 1003, or installed from the ROM 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the methods of the embodiments disclosed in the present application are executed.

[0155] The sonar image semantic communication device provided by this application adopts the sonar image semantic communication method in the above-mentioned embodiment, which can solve the technical problems of limited bandwidth of the link channel, dynamic change of channel quality, and poor communication quality when transmitting far-sea ocean data from the sea surface to the satellite and then to the land through a low-earth orbit satellite as a relay for a large amount of ocean data such as sonar images or videos. Compared with the prior art, the beneficial effects of the sonar image semantic communication device provided by this application are the same as those of the sonar image semantic communication method provided by the above-mentioned embodiment, and other technical features in this sonar image semantic communication device are the same as those disclosed in the method of the previous embodiment, which will not be elaborated here.

[0156] It should be understood that each part disclosed in this application can be implemented by hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in a suitable manner in any one or more embodiments or examples.

[0157] The above is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art within the technical scope disclosed in this application can easily think of changes or substitutions, which should all be covered by the protection scope of this application. Therefore, the protection scope of this application should be subject to the protection scope of the claims.

[0158] This application provides a computer-readable storage medium with computer-readable program instructions (i.e., computer programs) stored thereon, and the computer-readable program instructions are used to execute the sonar image semantic communication method in the above-mentioned embodiment.

[0159] The computer-readable storage medium provided by the present application may be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or components, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections with one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM) or flash memory, optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the above. In this embodiment, the computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, device, or component. The program code contained on the computer-readable storage medium can be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0160] The above computer-readable storage medium may be included in the sonar image semantic communication device; or it may exist separately without being assembled into the sonar image semantic communication device.

[0161] The above computer-readable storage medium carries one or more programs, and when the one or more programs are executed by the sonar image semantic communication device, the sonar image semantic communication device is caused to execute the sonar image semantic communication method described above.

[0162] Computer program code for performing the operations of the present application may be written in one or more programming languages or combinations thereof. The above programming languages include object-oriented programming languages - such as Java, Smalltalk, C++; and also include conventional procedural programming languages - such as the "C" language or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, executed as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network - including a local area network (LAN) or a wide area network (WAN) - or may be connected to an external computer (e.g., by connecting through the Internet using an Internet service provider).

[0163] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of the present application. In this regard, each block in the flowchart or block diagram may represent a module, a segment of a program, or a part of code that contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than that marked in the accompanying drawings. For example, two consecutive blocks shown may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and combinations of blocks in the block diagram and / or flowchart, may be implemented by a dedicated hardware-based system that performs the specified functions or operations, or may be implemented by a combination of dedicated hardware and computer instructions.

[0164] The modules described in the embodiments of the present application can be implemented in software or in hardware. In some cases, the name of the module does not constitute a limitation on the unit itself.

[0165] The readable storage medium provided by the present application is a computer-readable storage medium, and the computer-readable storage medium stores computer-readable program instructions (i.e., computer programs) for executing the above-mentioned sonar image semantic communication method, which can solve the technical problems of limited bandwidth of the link channel and dynamic change of channel quality and poor communication quality when transmitting far-sea marine data from the sea surface to the satellite and then to the land through a low-earth orbit satellite as a relay for a large amount of marine data such as sonar images or videos. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided by the present application are the same as those of the sonar image semantic communication method provided by the above embodiments, and will not be elaborated here.

[0166] The present application also provides a computer program product, including a computer program, and when the computer program is executed by a processor, it implements the steps of the sonar image semantic communication method as described above.

[0167] The computer program product provided by the present application can solve the technical problems of limited bandwidth of the link channel and dynamic change of channel quality and poor communication quality when transmitting far-sea marine data from the sea surface to the satellite and then to the land through a low-earth orbit satellite as a relay for a large amount of marine data such as sonar images or videos. Compared with the prior art, the beneficial effects of the computer program product provided by the present application are the same as those of the sonar image semantic communication method provided by the above embodiments, and will not be elaborated here.

[0168] The above are only some embodiments of the present application, and do not limit the patent scope of the present application. Any equivalent structural transformation made under the technical concept of the present application by using the content of the specification and drawings of the present application, or direct / indirect application in other related technical fields, is included in the patent protection scope of the present application.

Claims

1. A sonar image semantic communication method, characterized in that: The sonar image semantic communication method is applied to a network sending end, comprising: Acquire an original sonar image, and perform semantic encoding on the original sonar image according to a semantic network to extract semantic features; Combining the semantic features with the signal-to-noise ratio information of the channel to obtain channel coding output vector data; The channel coding output vector data is compressed based on the quantization module to obtain the target integer data, and the target integer data is sent to the network receiving end through the channel. The quantization function of the quantization module is: ,in, is the low-dimensional semantics output by the semantic encoder, is the quantitative result, is the scale parameter to be learned, is the rounding function; The network receiving end restores the received target integer data to the channel coding output vector data based on the quantization factor, and performs channel decoding and semantic decoding on the channel coding output vector data to obtain the target sonar image. The inverse quantization function corresponding to the quantization factor is: ,in To restore the vector to floating point data, is a bias term to eliminate part of the additive noise; The step of obtaining the original sonar image, and semantically encoding the original sonar image according to the semantic network to extract semantic features includes: Obtain original sonar images collected by marine operating equipment; Dividing the original sonar image into blocks according to a network hyperparameter setting table preset in the semantic network to obtain small-sized sonar image blocks; Transmitting the sonar image block to a multi-layer Swin Transformer module to extract the semantic features of the original sonar image; Before the step of obtaining the original sonar image collected by the marine operation equipment, the method further includes: Selecting a synthetic aperture radar data set to perform pre-training on the semantic network to obtain pre-training network parameters; Selecting a fine-tuning method based on transfer learning, adjusting the pre-trained network parameters using a preset sonar image data set to obtain target network parameters; The target network parameters are applied in the semantic network to construct the network hyperparameter setting table and the SwinTransformer module.

2. The sonar image semantic communication method according to claim 1, characterized in that: The step of transmitting the sonar image block to a multi-layer Swin Transformer module to extract the semantic features of the original sonar image comprises: The divided sonar image blocks are sent as input to the multi-layer Swin Transformer module; In the multi-layer Swin Transformer module, a fixed-window multi-head attention mechanism and a moving-window multi-head attention mechanism are alternately used to capture the local and global initial semantic features of the sonar image block; The initial semantic features are calculated and adjusted through residual connection and layer normalization, and the adjusted initial semantic features are integrated to obtain the semantic features of the original sonar image.

3. The sonar image semantic communication method according to claim 1, characterized in that: The step of combining the semantic feature with the signal-to-noise ratio information of the channel to obtain channel coding output vector data comprises: Collect the signal-to-noise ratio information of the current channel; Obtaining global statistical information of the semantic feature on the entire feature map based on a global average pooling function, and concatenating the global statistical information with the signal-to-noise ratio information to obtain context information; The context information is fed into a fully connected neural network, and a set of weight factors is calculated based on the semantic features based on a signal-to-noise ratio attention module; The semantic features are multiplied by the corresponding weight factors respectively to obtain channel coding output vector data.

4. A sonar image semantic communication device, characterized in that: The device comprises: A feature extraction module is used to obtain an original sonar image, and to perform semantic encoding on the original sonar image according to a semantic network to extract semantic features; A data generation module, used for combining the semantic features with the signal-to-noise ratio information of the channel to obtain channel coding output vector data; A data transmission module is used to compress the channel coding output vector data based on the quantization module to obtain target integer data, and send the target integer data to the network receiving end through the channel. The quantization function of the quantization module is ,in, is the low-dimensional semantics output by the semantic encoder, is the quantitative result, is the scale parameter to be learned, is the rounding function; The image reconstruction module is used to restore the received integer data to the channel coding output vector data based on the quantization factor, and perform channel decoding and semantic decoding on the channel coding output vector data to obtain the target sonar image. The inverse quantization function corresponding to the quantization factor is: ,in To restore the vector to floating point data, is a bias term to eliminate part of the additive noise; The feature extraction module is also used to obtain the original sonar image collected by the marine operation equipment; divide the original sonar image into blocks according to the network hyperparameter setting table preset in the semantic network to obtain small-sized sonar image blocks; and transmit the sonar image blocks to the multi-layer Swin Transformer module to extract the semantic features of the original sonar image; The sonar image semantic communication device is also used to select a synthetic aperture radar data set to pre-train on the semantic network to obtain pre-trained network parameters; based on the transfer learning, a fine-tuning method is selected to adjust the pre-trained network parameters using a preset sonar image data set to obtain target network parameters; the target network parameters are applied to the semantic network to construct the network hyperparameter setting table and the Swin Transformer module.

5. A sonar image semantic communication device, characterized in that: The device comprises: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the sonar image semantic communication method according to any one of claims 1 to 3.

6. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the sonar image semantic communication method according to any one of claims 1 to 3 are implemented.

7. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the steps of the sonar image semantic communication method according to any one of claims 1 to 3 are implemented.

Images