Some statistical characteristics of signal and noise in shallow-water acoustical channel were analysed . Based on the differences in some statistical characteristics between the signal and the noise , the authors deve...Some statistical characteristics of signal and noise in shallow-water acoustical channel were analysed . Based on the differences in some statistical characteristics between the signal and the noise , the authors developed a new kind of signal processing technique-digital time correlative accumulation to operate sonar systems at a low data rate . Theoretical analyses and expermental results show . the false-alarm probability can be reduced to a low value of less than 10 -4 while the detection probability can reach a relatively high value of more than 0.9. By using the statistical averages of the multiple range detection value , the authors can greatly improve the accuracy of the detection . The method can be used for some other related fields, for example , ultrasonic detection in air medium .展开更多
A new algorithm based on a Supervised Self-Organizing neural network for the pas sive sonar target recognition was proposed. Because of the incompleteness of the passive sonar exemplar set, the algorithm introduced a ...A new algorithm based on a Supervised Self-Organizing neural network for the pas sive sonar target recognition was proposed. Because of the incompleteness of the passive sonar exemplar set, the algorithm introduced a Multi-Activation-function structure and Supervised Self-Organizing competitive learning algorithm into the classic feed-forward neural networks,and obviously improved the generalization ability in target recognition. Besides, it can effi ciently reduce the learning time and avoid the local optimum. The recognition experiments of realistic passive sonar signals show that this new algorithm has good generalization ability and high recognition展开更多
Underwater target localization and parameters(azimuth and range) estimation by the method of utilizing explosions as underwater sound sources are described in this paper.The narrow beam reverberation model of the targ...Underwater target localization and parameters(azimuth and range) estimation by the method of utilizing explosions as underwater sound sources are described in this paper.The narrow beam reverberation model of the target echo signal is researched to estimate the target azimuth in reverberation background.Estimation errors of target azimuth and range are studied and proved to approximately meet Gauss distribution.Then the variance formula of target range error is deduced.Simulation experiments are applied to research the target range error and its standard deviation,and a series of measures to improve the estimation accuracy of target range are proposed.It is confirmed by the data processing results of simulations and lake experiments that the proposed method can accurately locate underwater target at a long distance on the condition of a certain underwater explosion range error.展开更多
In order to study the optimization configuration problem of bistatic sonar system,the optimal configuration model of bistatic sonar system was established.The positioning accuracy of the dual-base sonar at different d...In order to study the optimization configuration problem of bistatic sonar system,the optimal configuration model of bistatic sonar system was established.The positioning accuracy of the dual-base sonar at different dual-base angles was obtained by calculating the Cramér-Rao lower bound(CRLB)accuracy based on the optimal configuration model.The effect of system time measurement error and angle measurement error on positioning accuracy was analyzed by simulation.The transmitting and receiving sonar are deployed on the same circle centered on the target in the simulation.The results showed that the positioning accuracy was the highest at the bistatic angle of 2π/3.This research had certain reference for the optimal configuration of bistatic/multistatic sonar system.展开更多
Sonar image processing system is an important intelligent system of Autonomous Un-derwater Vehicle.Based on TMS320C30 high speed DSP,it is used to realize sonar imagecompression and underwater object detections includ...Sonar image processing system is an important intelligent system of Autonomous Un-derwater Vehicle.Based on TMS320C30 high speed DSP,it is used to realize sonar imagecompression and underwater object detections including obstacle recognition in real time.Inthis paper,the software and hardware designs of this system are introduced and the experi-mental results are given.展开更多
Based on the principle of active sonar detection,a pipeline sonar sys-tem has been developed for detecting the water accumulated in undergroundgas-pipelines,and the effectiveness of this system verified through field-...Based on the principle of active sonar detection,a pipeline sonar sys-tem has been developed for detecting the water accumulated in undergroundgas-pipelines,and the effectiveness of this system verified through field-testing.The working process,experimental results and some considerations in the deter-mination of sonar parameters are described in this paper.展开更多
Seabed sediment recognition is vital for the exploitation of marine resources.Side-scan sonar(SSS)is an excellent tool for acquiring the imagery of seafloor topography.Combined with ocean surface sampling,it provides ...Seabed sediment recognition is vital for the exploitation of marine resources.Side-scan sonar(SSS)is an excellent tool for acquiring the imagery of seafloor topography.Combined with ocean surface sampling,it provides detailed and accurate images of marine substrate features.Most of the processing of SSS imagery works around limited sampling stations and requires manual interpretation to complete the classification of seabed sediment imagery.In complex sea areas,with manual interpretation,small targets are often lost due to a large amount of information.To date,studies related to the automatic recognition of seabed sediments are still few.This paper proposes a seabed sediment recognition method based on You Only Look Once version 5 and SSS imagery to perform real-time sedi-ment classification and localization for accuracy,particularly on small targets and faster speeds.We used methods such as changing the dataset size,epoch,and optimizer and adding multiscale training to overcome the challenges of having a small sample and a low accuracy.With these methods,we improved the results on mean average precision by 8.98%and F1 score by 11.12%compared with the original method.In addition,the detection speed was approximately 100 frames per second,which is faster than that of previous methods.This speed enabled us to achieve real-time seabed sediment recognition from SSS imagery.展开更多
Side scan sonar(SSS)is an important means to detect and locate seafloor targets.Autonomous underwater vehicles(AUVs)carrying SSS stay near the seafloor to obtain high-resolution images and provide the outline of the t...Side scan sonar(SSS)is an important means to detect and locate seafloor targets.Autonomous underwater vehicles(AUVs)carrying SSS stay near the seafloor to obtain high-resolution images and provide the outline of the target for observers.The target feature information of an SSS image is similar to the background information,and a small target has less pixel information;therefore,accu-rately identifying and locating small targets in SSS images is challenging.We collect the SSS images of iron metal balls(with a diameter of 1m)and rocks to solve the problem of target misclassification.Thus,the dataset contains two types of targets,namely,‘ball’and‘rock’.With the aim to enable AUVs to accurately and automatically identify small underwater targets in SSS images,this study designs a multisize parallel convolution module embedded in state-of-the-art Yolo5.An attention mechanism transformer and a convolutional block attention module are also introduced to compare their contributions to small target detection accuracy.The performance of the proposed method is further evaluated by taking the lightweight networks Mobilenet3 and Shufflenet2 as the backbone network of Yolo5.This study focuses on the performance of convolutional neural networks for the detection of small targets in SSS images,while another comparison experiment is carried out using traditional HOG+SVM to highlight the neural network’s ability.This study aims to improve the detection accuracy while ensuring the model efficiency to meet the real-time working requirements of AUV target detection.展开更多
It has remained a hard nut for years to segment sonar images of jacket installation environment,most of which are noisy images with inevitable blur after noise reduction.For the purpose of solutions to this problem,a ...It has remained a hard nut for years to segment sonar images of jacket installation environment,most of which are noisy images with inevitable blur after noise reduction.For the purpose of solutions to this problem,a fast segmen-tation algorithm is proposed on the basis of the gray value characteristics of sonar images.This algorithm is endowed with the advantage in no need of segmentation thresholds.To realize this goal,we follow the undermentioned steps:first,calcu-late the gray matrix of the fuzzy image background.After adjusting the gray value,the image is divided into three regions:background region,buffer region and target regions.Afterfiltering,we reset the pixels with gray value lower than 255 to binarize images and eliminate most artifacts.Finally,the remaining noise is removed by morphological processing.The simulation results of several sonar images show that the algorithm can segment the fuzzy sonar images quickly and effectively.Thus,the stable and feasible method is testified.展开更多
Karst landforms are widely distributed in China,and are most common in Yunnan,Guizhou and Guangxi.If the development of karst caves at the bottom of the piles cannot be accurately ascertained before the construction o...Karst landforms are widely distributed in China,and are most common in Yunnan,Guizhou and Guangxi.If the development of karst caves at the bottom of the piles cannot be accurately ascertained before the construction of bridge pile foundations,accidents such as hole collapse,slurry leakage,and drill sticking will easily occur.In this paper,the principle and method of sonar detection for detecting karst caves at the bottom of bridge piles was introduced,and the sonar detection data and the cave situation at the bottom of the pile during the construction process in combination with the case of Yunnan Zhenguo Highway Project was analyzed,which verifies the practicability and reliability of sonar detection method reliability.展开更多
Understanding the topographic patterns of the seafloor is a very important part of understanding our planet.Although the science involved in bathymetric surveying has advanced much over the decades,less than 20%of the...Understanding the topographic patterns of the seafloor is a very important part of understanding our planet.Although the science involved in bathymetric surveying has advanced much over the decades,less than 20%of the seafloor has been precisely modeled to date,and there is an urgent need to improve the accuracy and reduce the uncertainty of underwater survey data.In this study,we introduce a pretrained visual geometry group network(VGGNet)method based on deep learning.To apply this method,we input gravity anomaly data derived from ship measurements and satellite altimetry into the model and correct the latter,which has a larger spatial coverage,based on the former,which is considered the true value and is more accurate.After obtaining the corrected high-precision gravity model,it is inverted to the corresponding bathymetric model by applying the gravity-depth correlation.We choose four data pairs collected from different environments,i.e.,the Southern Ocean,Pacific Ocean,Atlantic Ocean and Caribbean Sea,to evaluate the topographic correction results of the model.The experiments show that the coefficient of determination(R~2)reaches 0.834 among the results of the four experimental groups,signifying a high correlation.The standard deviation and normalized root mean square error are also evaluated,and the accuracy of their performance improved by up to 24.2%compared with similar research done in recent years.The evaluation of the R^(2) values at different water depths shows that our model can achieve performance results above 0.90 at certain water depths and can also significantly improve results from mid-water depths when compared to previous research.Finally,the bathymetry corrected by our model is able to show an accuracy improvement level of more than 21%within 1%of the total water depths,which is sufficient to prove that the VGGNet-based method has the ability to perform a gravity-bathymetry correction and achieve outstanding results.展开更多
Traditional direction of arrival(DOA)estimation methods based on sparse reconstruction commonly use convex or smooth functions to approximate non-convex and non-smooth sparse representation problems.This approach ofte...Traditional direction of arrival(DOA)estimation methods based on sparse reconstruction commonly use convex or smooth functions to approximate non-convex and non-smooth sparse representation problems.This approach often introduces errors into the sparse representation model,necessitating the development of improved DOA estimation algorithms.Moreover,conventional DOA estimation methods typically assume that the signal coincides with a predetermined grid.However,in reality,this assumption often does not hold true.The likelihood of a signal not aligning precisely with the predefined grid is high,resulting in potential grid mismatch issues for the algorithm.To address the challenges associated with grid mismatch and errors in sparse representation models,this article proposes a novel high-performance off-grid DOA estimation approach based on iterative proximal projection(IPP).In the proposed method,we employ an alternating optimization strategy to jointly estimate sparse signals and grid offset parameters.A proximal function optimization model is utilized to address non-convex and non-smooth sparse representation problems in DOA estimation.Subsequently,we leverage the smoothly clipped absolute deviation penalty(SCAD)function to compute the proximal operator for solving the model.Simulation and sea trial experiments have validated the superiority of the proposed method in terms of higher resolution and more accurate DOA estimation performance when compared to both traditional sparse reconstruction methods and advanced off-grid techniques.展开更多
With the development of underwater sonar detection technology,simultaneous localization and mapping(SLAM)approach has attracted much attention in underwater navigation field in recent years.But the weak detection abil...With the development of underwater sonar detection technology,simultaneous localization and mapping(SLAM)approach has attracted much attention in underwater navigation field in recent years.But the weak detection ability of a single vehicle limits the SLAM performance in wide areas.Thereby,cooperative SLAM using multiple vehicles has become an important research direction.The key factor of cooperative SLAM is timely and efficient sonar image transmission among underwater vehicles.However,the limited bandwidth of underwater acoustic channels contradicts a large amount of sonar image data.It is essential to compress the images before transmission.Recently,deep neural networks have great value in image compression by virtue of the powerful learning ability of neural networks,but the existing sonar image compression methods based on neural network usually focus on the pixel-level information without the semantic-level information.In this paper,we propose a novel underwater acoustic transmission scheme called UAT-SSIC that includes semantic segmentation-based sonar image compression(SSIC)framework and the joint source-channel codec,to improve the accuracy of the semantic information of the reconstructed sonar image at the receiver.The SSIC framework consists of Auto-Encoder structure-based sonar image compression network,which is measured by a semantic segmentation network's residual.Considering that sonar images have the characteristics of blurred target edges,the semantic segmentation network used a special dilated convolution neural network(DiCNN)to enhance segmentation accuracy by expanding the range of receptive fields.The joint source-channel codec with unequal error protection is proposed that adjusts the power level of the transmitted data,which deal with sonar image transmission error caused by the serious underwater acoustic channel.Experiment results demonstrate that our method preserves more semantic information,with advantages over existing methods at the same compression ratio.It also improves the error tolerance and packet loss resistance of transmission.展开更多
文摘Some statistical characteristics of signal and noise in shallow-water acoustical channel were analysed . Based on the differences in some statistical characteristics between the signal and the noise , the authors developed a new kind of signal processing technique-digital time correlative accumulation to operate sonar systems at a low data rate . Theoretical analyses and expermental results show . the false-alarm probability can be reduced to a low value of less than 10 -4 while the detection probability can reach a relatively high value of more than 0.9. By using the statistical averages of the multiple range detection value , the authors can greatly improve the accuracy of the detection . The method can be used for some other related fields, for example , ultrasonic detection in air medium .
文摘A new algorithm based on a Supervised Self-Organizing neural network for the pas sive sonar target recognition was proposed. Because of the incompleteness of the passive sonar exemplar set, the algorithm introduced a Multi-Activation-function structure and Supervised Self-Organizing competitive learning algorithm into the classic feed-forward neural networks,and obviously improved the generalization ability in target recognition. Besides, it can effi ciently reduce the learning time and avoid the local optimum. The recognition experiments of realistic passive sonar signals show that this new algorithm has good generalization ability and high recognition
基金supported by the National Natural Science Foundation of China(61431020,61571434)
文摘Underwater target localization and parameters(azimuth and range) estimation by the method of utilizing explosions as underwater sound sources are described in this paper.The narrow beam reverberation model of the target echo signal is researched to estimate the target azimuth in reverberation background.Estimation errors of target azimuth and range are studied and proved to approximately meet Gauss distribution.Then the variance formula of target range error is deduced.Simulation experiments are applied to research the target range error and its standard deviation,and a series of measures to improve the estimation accuracy of target range are proposed.It is confirmed by the data processing results of simulations and lake experiments that the proposed method can accurately locate underwater target at a long distance on the condition of a certain underwater explosion range error.
基金supported by the National High Technology Research and Development Program of China(863 Program,18-H863-04-2T-003-019-01)。
文摘In order to study the optimization configuration problem of bistatic sonar system,the optimal configuration model of bistatic sonar system was established.The positioning accuracy of the dual-base sonar at different dual-base angles was obtained by calculating the Cramér-Rao lower bound(CRLB)accuracy based on the optimal configuration model.The effect of system time measurement error and angle measurement error on positioning accuracy was analyzed by simulation.The transmitting and receiving sonar are deployed on the same circle centered on the target in the simulation.The results showed that the positioning accuracy was the highest at the bistatic angle of 2π/3.This research had certain reference for the optimal configuration of bistatic/multistatic sonar system.
基金the High Technology Research and Development Programme of china.
文摘Sonar image processing system is an important intelligent system of Autonomous Un-derwater Vehicle.Based on TMS320C30 high speed DSP,it is used to realize sonar imagecompression and underwater object detections including obstacle recognition in real time.Inthis paper,the software and hardware designs of this system are introduced and the experi-mental results are given.
文摘Based on the principle of active sonar detection,a pipeline sonar sys-tem has been developed for detecting the water accumulated in undergroundgas-pipelines,and the effectiveness of this system verified through field-testing.The working process,experimental results and some considerations in the deter-mination of sonar parameters are described in this paper.
基金funded by the Natural Science Foundation of Fujian Province(No.2018J01063)the Project of Deep Learning Based Underwater Cultural Relics Recognization(No.38360041)the Project of the State Administration of Cultural Relics(No.2018300).
文摘Seabed sediment recognition is vital for the exploitation of marine resources.Side-scan sonar(SSS)is an excellent tool for acquiring the imagery of seafloor topography.Combined with ocean surface sampling,it provides detailed and accurate images of marine substrate features.Most of the processing of SSS imagery works around limited sampling stations and requires manual interpretation to complete the classification of seabed sediment imagery.In complex sea areas,with manual interpretation,small targets are often lost due to a large amount of information.To date,studies related to the automatic recognition of seabed sediments are still few.This paper proposes a seabed sediment recognition method based on You Only Look Once version 5 and SSS imagery to perform real-time sedi-ment classification and localization for accuracy,particularly on small targets and faster speeds.We used methods such as changing the dataset size,epoch,and optimizer and adding multiscale training to overcome the challenges of having a small sample and a low accuracy.With these methods,we improved the results on mean average precision by 8.98%and F1 score by 11.12%compared with the original method.In addition,the detection speed was approximately 100 frames per second,which is faster than that of previous methods.This speed enabled us to achieve real-time seabed sediment recognition from SSS imagery.
基金supported by the National Key Research and Development Program of China(No.2016YFC0301400).
文摘Side scan sonar(SSS)is an important means to detect and locate seafloor targets.Autonomous underwater vehicles(AUVs)carrying SSS stay near the seafloor to obtain high-resolution images and provide the outline of the target for observers.The target feature information of an SSS image is similar to the background information,and a small target has less pixel information;therefore,accu-rately identifying and locating small targets in SSS images is challenging.We collect the SSS images of iron metal balls(with a diameter of 1m)and rocks to solve the problem of target misclassification.Thus,the dataset contains two types of targets,namely,‘ball’and‘rock’.With the aim to enable AUVs to accurately and automatically identify small underwater targets in SSS images,this study designs a multisize parallel convolution module embedded in state-of-the-art Yolo5.An attention mechanism transformer and a convolutional block attention module are also introduced to compare their contributions to small target detection accuracy.The performance of the proposed method is further evaluated by taking the lightweight networks Mobilenet3 and Shufflenet2 as the backbone network of Yolo5.This study focuses on the performance of convolutional neural networks for the detection of small targets in SSS images,while another comparison experiment is carried out using traditional HOG+SVM to highlight the neural network’s ability.This study aims to improve the detection accuracy while ensuring the model efficiency to meet the real-time working requirements of AUV target detection.
基金supported by Open Fund Project of China Key Laboratory of Submarine Geoscience(KLSG1802)Science&Technology Project of China Ocean Mineral Resources Research and Development Association(DY135-N1-1-05)Science&Technology Project of Zhoushan city of Zhejiang Province(2019C42271,2019C33205).
文摘It has remained a hard nut for years to segment sonar images of jacket installation environment,most of which are noisy images with inevitable blur after noise reduction.For the purpose of solutions to this problem,a fast segmen-tation algorithm is proposed on the basis of the gray value characteristics of sonar images.This algorithm is endowed with the advantage in no need of segmentation thresholds.To realize this goal,we follow the undermentioned steps:first,calcu-late the gray matrix of the fuzzy image background.After adjusting the gray value,the image is divided into three regions:background region,buffer region and target regions.Afterfiltering,we reset the pixels with gray value lower than 255 to binarize images and eliminate most artifacts.Finally,the remaining noise is removed by morphological processing.The simulation results of several sonar images show that the algorithm can segment the fuzzy sonar images quickly and effectively.Thus,the stable and feasible method is testified.
文摘Karst landforms are widely distributed in China,and are most common in Yunnan,Guizhou and Guangxi.If the development of karst caves at the bottom of the piles cannot be accurately ascertained before the construction of bridge pile foundations,accidents such as hole collapse,slurry leakage,and drill sticking will easily occur.In this paper,the principle and method of sonar detection for detecting karst caves at the bottom of bridge piles was introduced,and the sonar detection data and the cave situation at the bottom of the pile during the construction process in combination with the case of Yunnan Zhenguo Highway Project was analyzed,which verifies the practicability and reliability of sonar detection method reliability.
基金The National Key R&D Program of China under contract Nos 2022YFC3003800,2020YFC1521700 and 2020YFC1521705the National Natural Science Foundation of China under contract No.41830540+3 种基金the Open Fund of the East China Coastal Field Scientific Observation and Research Station of the Ministry of Natural Resources under contract No.OR-SECCZ2022104the Deep Blue Project of Shanghai Jiao Tong University under contract No.SL2020ZD204the Special Funding Project for the Basic Scientific Research Operation Expenses of the Central Government-Level Research Institutes of Public Interest of China under contract No.SZ2102the Zhejiang Provincial Project under contract No.330000210130313013006。
文摘Understanding the topographic patterns of the seafloor is a very important part of understanding our planet.Although the science involved in bathymetric surveying has advanced much over the decades,less than 20%of the seafloor has been precisely modeled to date,and there is an urgent need to improve the accuracy and reduce the uncertainty of underwater survey data.In this study,we introduce a pretrained visual geometry group network(VGGNet)method based on deep learning.To apply this method,we input gravity anomaly data derived from ship measurements and satellite altimetry into the model and correct the latter,which has a larger spatial coverage,based on the former,which is considered the true value and is more accurate.After obtaining the corrected high-precision gravity model,it is inverted to the corresponding bathymetric model by applying the gravity-depth correlation.We choose four data pairs collected from different environments,i.e.,the Southern Ocean,Pacific Ocean,Atlantic Ocean and Caribbean Sea,to evaluate the topographic correction results of the model.The experiments show that the coefficient of determination(R~2)reaches 0.834 among the results of the four experimental groups,signifying a high correlation.The standard deviation and normalized root mean square error are also evaluated,and the accuracy of their performance improved by up to 24.2%compared with similar research done in recent years.The evaluation of the R^(2) values at different water depths shows that our model can achieve performance results above 0.90 at certain water depths and can also significantly improve results from mid-water depths when compared to previous research.Finally,the bathymetry corrected by our model is able to show an accuracy improvement level of more than 21%within 1%of the total water depths,which is sufficient to prove that the VGGNet-based method has the ability to perform a gravity-bathymetry correction and achieve outstanding results.
基金supported by the National Science Foundation for Distinguished Young Scholars(Grant No.62125104)the National Natural Science Foundation of China(Grant No.52071111).
文摘Traditional direction of arrival(DOA)estimation methods based on sparse reconstruction commonly use convex or smooth functions to approximate non-convex and non-smooth sparse representation problems.This approach often introduces errors into the sparse representation model,necessitating the development of improved DOA estimation algorithms.Moreover,conventional DOA estimation methods typically assume that the signal coincides with a predetermined grid.However,in reality,this assumption often does not hold true.The likelihood of a signal not aligning precisely with the predefined grid is high,resulting in potential grid mismatch issues for the algorithm.To address the challenges associated with grid mismatch and errors in sparse representation models,this article proposes a novel high-performance off-grid DOA estimation approach based on iterative proximal projection(IPP).In the proposed method,we employ an alternating optimization strategy to jointly estimate sparse signals and grid offset parameters.A proximal function optimization model is utilized to address non-convex and non-smooth sparse representation problems in DOA estimation.Subsequently,we leverage the smoothly clipped absolute deviation penalty(SCAD)function to compute the proximal operator for solving the model.Simulation and sea trial experiments have validated the superiority of the proposed method in terms of higher resolution and more accurate DOA estimation performance when compared to both traditional sparse reconstruction methods and advanced off-grid techniques.
基金supported in part by the Tianjin Technology Innovation Guidance Special Fund Project under Grant No.21YDTPJC00850in part by the National Natural Science Foundation of China under Grant No.41906161in part by the Natural Science Foundation of Tianjin under Grant No.21JCQNJC00650。
文摘With the development of underwater sonar detection technology,simultaneous localization and mapping(SLAM)approach has attracted much attention in underwater navigation field in recent years.But the weak detection ability of a single vehicle limits the SLAM performance in wide areas.Thereby,cooperative SLAM using multiple vehicles has become an important research direction.The key factor of cooperative SLAM is timely and efficient sonar image transmission among underwater vehicles.However,the limited bandwidth of underwater acoustic channels contradicts a large amount of sonar image data.It is essential to compress the images before transmission.Recently,deep neural networks have great value in image compression by virtue of the powerful learning ability of neural networks,but the existing sonar image compression methods based on neural network usually focus on the pixel-level information without the semantic-level information.In this paper,we propose a novel underwater acoustic transmission scheme called UAT-SSIC that includes semantic segmentation-based sonar image compression(SSIC)framework and the joint source-channel codec,to improve the accuracy of the semantic information of the reconstructed sonar image at the receiver.The SSIC framework consists of Auto-Encoder structure-based sonar image compression network,which is measured by a semantic segmentation network's residual.Considering that sonar images have the characteristics of blurred target edges,the semantic segmentation network used a special dilated convolution neural network(DiCNN)to enhance segmentation accuracy by expanding the range of receptive fields.The joint source-channel codec with unequal error protection is proposed that adjusts the power level of the transmitted data,which deal with sonar image transmission error caused by the serious underwater acoustic channel.Experiment results demonstrate that our method preserves more semantic information,with advantages over existing methods at the same compression ratio.It also improves the error tolerance and packet loss resistance of transmission.