AquaPile-YOLO: Pioneering Underwater Pile Foundation Detection with Forward-Looking Sonar Image Processing

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Comments：

This manuscript mainly presented a new method, AquaPile-Yolo, to detect underwater pile foundation based on sonar image process. The results showed that this method can achieve good performance. This work holds both scientific and practical significance. The logic is very clear, and the language is also well organized. I suggest it can be accepted in Remote Sensing after some minor revisions.

1 Major problems

1)Study scope

The journal Remote Sensing primarily focuses on methods and applications for satellite remote sensing. However, this manuscript primarily presents a new underwater image processing method. The authors should describe why it is suitable for this journal using more words.

2)Introduction

The introduction should put the research in the context, introduce the related previous studies, and propose a proper scientific question. It seems that the second part of this manuscript, Related Work, also falls within the scope of Introduction. It is recommeded to merge the first two parts and shorten the text to make the article more concise.

3)Discussion and Conclusions

The manuscript could benefit from a reorganization of the "Discussion" and "Conclusions" sections. Specifically, some of the content discussing future research directions would be more appropriately placed in the "Discussion" section. The "Conclusions" should focus on summarizing the main findings and contributions of the study.

Some forward-looking statements in the conclusions are more appropriate for the discussion section. For example, aspects related to potential applications in autonomous underwater vehicles (AUVs) and environmental monitoring could be detailed in "Discussion."

Strengthen the conclusion by focusing on quantitative achievements and the practical implications of the results.

While the manuscript outlines future research directions, including a detailed roadmap for implementing these advancements, such as multi-modal data fusion or deployment in resource-constrained environments, would be beneficial.

2 some minor problems

For better visual organization, figures of the same type or category could be combined where appropriate. For instance, Figure 4 and Figure 5, which depict experimental instruments, could be merged into a single figure panel. Similarly, the experimental results in Figure 13 and Figure 14 could be combined to provide a more concise representation of the findings.

Author Response

Comments 1: Study scope

The journal Remote Sensing primarily focuses on methods and applications for satellite remote sensing. However, this manuscript primarily presents a new underwater image processing method. The authors should describe why it is suitable for this journal using more words.

Response 1: While this manuscript focuses on underwater sonar image processing, we believe it aligns well with the scope of Remote Sensing. Sonar imaging, though not directly utilizing satellite-based remote sensing, shares significant similarities with satellite imaging in terms of data processing methodologies, AI-driven target recognition, and feature extraction techniques.

Moreover, this work directly contributes to the journal's Special Issue on "Artificial Intelligence for Ocean Remote Sensing." The integration of artificial intelligence in this study demonstrates the potential of advanced machine learning algorithms for solving complex oceanic problems, akin to the challenges addressed in satellite-based remote sensing. By leveraging AI for underwater sonar imagery, this study supports a broader understanding of ocean environments, complementing large-scale satellite observations and advancing data-driven paradigms in oceanography.

Comments 2: Introduction

The introduction should put the research in the context, introduce the related previous studies, and propose a proper scientific question. It seems that the second part of this manuscript, Related Work, also falls within the scope of Introduction. It is recommeded to merge the first two parts and shorten the text to make the article more concise.

Response 2: In response to your suggestion, we have merged the "Introduction" and "Related Work" sections into a single cohesive section. During this process, we reorganized the paragraphs to establish a clearer logical flow, ensuring that the research context, relevant prior studies, and the proposed scientific question are all presented succinctly and effectively. Additionally, we revised and streamlined the content to make the article more concise, aligning it with the journal's standards for clarity and readability.

Comments 3: Discussion and Conclusions

The manuscript could benefit from a reorganization of the "Discussion" and "Conclusions" sections. Specifically, some of the content discussing future research directions would be more appropriately placed in the "Discussion" section. The "Conclusions" should focus on summarizing the main findings and contributions of the study.

Some forward-looking statements in the conclusions are more appropriate for the discussion section. For example, aspects related to potential applications in autonomous underwater vehicles (AUVs) and environmental monitoring could be detailed in "Discussion."

Strengthen the conclusion by focusing on quantitative achievements and the practical implications of the results.

While the manuscript outlines future research directions, including a detailed roadmap for implementing these advancements, such as multi-modal data fusion or deployment in resource-constrained environments, would be beneficial.

Response 3:In response, we have made substantial revisions to the manuscript to improve the organization and focus of the "Discussion" and "Conclusions" sections. Specifically:

1.Reorganization of Sections:

We have integrated forward-looking statements and discussions about potential applications, such as the use of AquaPile-YOLO in autonomous underwater vehicles (AUVs) and environmental monitoring, into the "Discussion" section. This new section provides a detailed analysis of the study's implications, strengths, and limitations, as well as an outline of future research directions.

2.Strengthened Conclusions:

The "Conclusions" section has been revised to focus exclusively on summarizing the main findings and contributions of the study. We now emphasize quantitative achievements, such as the improved detection accuracy and robustness of AquaPile-YOLO, and discuss the practical implications of these results for real-world underwater applications.

3.Future Research Directions:

In the "Discussion" section, we outline a detailed roadmap for future advancements, including algorithm optimization for resource-constrained environments, multimodal data fusion, and broader environmental adaptability. These additions aim to address the challenges and expand the applicability of the proposed method.

Comments 4: some minor problems

For better visual organization, figures of the same type or category could be combined where appropriate. For instance, Figure 4 and Figure 5, which depict experimental instruments, could be merged into a single figure panel. Similarly, the experimental results in Figure 13 and Figure 14 could be combined to provide a more concise representation of the findings.

Response 4: In response,we have carefully considered your suggestion to combine figures of similar types or categories for enhanced visual clarity. Based on your feedback, we have made the following adjustments:

1.Combining Figures 4 and 5: We have merged the original Figures 4 and 5 into a new Figure 4. This consolidated figure now presents both the experimental instruments and their corresponding sonar data, providing a more coherent and comprehensive view of the experimental setup and its outputs.

2.Combining Figures 6 and 7: The original Figures 6 and 7 have been combined into a new Figure 5. This integration allows for a more compact presentation of related experimental results, facilitating easier comparison and analysis for readers.

3.Combining Figures 11 and 12: We have merged the original Figures 11 and 12 into a new Figure 9. This consolidation results in a more succinct presentation of the experimental findings, enabling readers to quickly grasp the key results of our study.

4.Combining Figures 13 and 14: The original Figures 13 and 14 have been combined into a new Figure 10. This new figure focuses on presenting the critical experimental data, allowing for a more intuitive understanding of the study's outcomes.

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

This manuscript presents a study on underwater pile foundation detection from sonar data. An AquaPile-YOLO method was proposed in this study by training on 4000 sonar images taken from a lake. Comparison between the proposed method and other methods were also presented, and results showed that the proposed AquaPile-YOLO method outperformed the others. This topic is quite interesting, and the method is quite useful in marine engineering, harbor operation, and ship navigation. This manuscript can be published after some improvements:

 

1. the introduction of this manuscript need to revise. Currently, the authors present an “introduction” and a “related work” section in the manuscript, actually these two sections can be merged into one as “introduction”. The logistic of the introduction section can be something like: general introduction of underwater pile foundation detection and its significance, followed by literature review of existing studies on this topic, then the drawbacks of these existing studies, challenges and knowledge gaps, finally the objectives of this study.

 

2. the discussion of this manuscript need to improve. Currently the discussion and results are in the same section, but the authors only presented results and there is no actual discussion. I would suggest the authors to add a separate section of discussion, and add some discussion points into this section, such as the advantages of this study, the limitation of this study, the applicability of the proposed method, future improvements.

 

3. I have some general questions. First, does wave influence the sonar collecting data as the sonar is installed 0.5m below the water surface? Second, what’s the depth range that the proposed method can work efficiently? Third, does the seabed types (e.g., muddy, rocky, sandy bottoms) matters? Some of these can be compiled into discussion.

 

Some other minor comments:

(1) line 298: please include the full name of C3 when it’s the first time mentioned in the manuscript.

(2) line 302: you have mentioned the full name of CBAM in line 297, so no need to mention again.

(3) line 382-386: please shorten this sentence, it’s repeated with line 153-156.

(4) the authors compared the proposed AquaPile-YOLO method with some other methods, are they trained using the same dataset?

(5) Figure 15: in the caption, please explain what does the red box represent.

Author Response

Comments 1: The introduction of this manuscript need to revise. Currently, the authors present an “introduction” and a “related work” section in the manuscript, actually these two sections can be merged into one as “introduction”. The logistic of the introduction section can be something like: general introduction of underwater pile foundation detection and its significance, followed by literature review of existing studies on this topic, then the drawbacks of these existing studies, challenges and knowledge gaps, finally the objectives of this study.

Response 1: In response to your suggestion, we have revised the introduction section by merging the "Introduction" and "Related Work" sections into a single, cohesive "Introduction" section. This new structure begins with a general introduction to the significance of underwater pile foundation detection, highlighting its importance in marine engineering and environmental monitoring. We then provide a comprehensive literature review of existing studies on this topic, discussing the advancements and methodologies that have been developed thus far. Following this, we identify the drawbacks of these existing studies, as well as the challenges and knowledge gaps that remain. Finally, we clearly outline the objectives of our study, which aim to address these gaps and improve upon the current state of underwater pile foundation detection technology.

Comments 2: The discussion of this manuscript need to improve. Currently the discussion and results are in the same section, but the authors only presented results and there is no actual discussion. I would suggest the authors to add a separate section of discussion, and add some discussion points into this section, such as the advantages of this study, the limitation of this study, the applicability of the proposed method, future improvements.

Response 2: In response to your suggestion, we have made significant revisions to our paper. Specifically, we have added a separate section titled "4.3. Discussion" to provide a comprehensive analysis of our findings. This new section includes a detailed discussion of the advantages of our study, such as the improved detection accuracy and robustness achieved by the AquaPile-YOLO algorithm. We also address the limitations of our study, particularly regarding the computational requirements and the need for further testing in diverse underwater environments.Furthermore, we discuss the applicability of the proposed method in marine engineering and environmental monitoring, highlighting its potential for enhancing the efficiency and safety of underwater operations. Additionally, we outline several avenues for future improvements, including algorithm optimization, multimodal data fusion, and broader environmental adaptability. To ensure a cohesive narrative, we have also integrated relevant content from the previous "Conclusions" section into the new "Discussion" section. We believe these changes will provide a more thorough and insightful analysis of our research, aligning with your expectations.

Comments 3: I have some general questions. First, does wave influence the sonar collecting data as the sonar is installed 0.5m below the water surface? Second, what’s the depth range that the proposed method can work efficiently? Third, does the seabed types (e.g., muddy, rocky, sandy bottoms) matters? Some of these can be compiled into discussion.

Response 3:

1.Wave Influence on Sonar Data Collection: While waves can potentially influence the sonar data collection process, the sonar is installed 0.5 meters below the water surface primarily to avoid the impact of surface bubbles, which can degrade the sonar's performance. Surface waves are indeed a significant cause of bubbles, and by positioning the sonar at this depth, we minimize the direct influence of these bubbles on the sonar images. In our experimental setup, this depth was found to be sufficient to maintain the integrity of the sonar data under normal conditions, ensuring that the sonar images were not adversely affected by surface disturbances. Nonetheless, we acknowledge that extreme wave conditions could still introduce some variability. To address this, we employed a stabilization system during data collection to further ensure the sonar maintained a consistent position relative to the target area. Additionally, we have discussed the potential effects of wave interference in the "Discussion" section of our revised manuscript, highlighting the need for further research to enhance the algorithm's robustness against such environmental factors.

2.Depth Range for Efficient Operation: The proposed AquaPile-YOLO method has been tested and demonstrated efficiency in a depth range of 2 to 20 meters, as utilized in our experimental setup. This range is suitable for many shallow water applications, such as harbor operations and coastal engineering projects. However, the depth range in which the method can operate efficiently may be influenced by factors such as sonar specifications, water quality, and target characteristics. We have noted the importance of exploring the method's performance across a broader depth range in future research to fully understand its capabilities and limitations.

3.Impact of Seabed Types: The seabed type, whether muddy, rocky, or sandy, can significantly affect the sonar imaging and target detection process. Different seabed types have varying acoustic properties, which can influence the reflection and scattering of sound waves, thereby impacting the quality and interpretation of the sonar images. In our study, we primarily focused on a substrate composed of sand and gravel. To address the influence of seabed types, we have included a discussion on this topic in the "Discussion" section, emphasizing the need for further investigation to assess the method's adaptability to diverse seabed conditions and to optimize the algorithm accordingly.

Comments 4: Some other minor comments:

(1) line 298: please include the full name of C3 when it’s the first time mentioned in the manuscript.

(2) line 302: you have mentioned the full name of CBAM in line 297, so no need to mention again.

(3) line 382-386: please shorten this sentence, it’s repeated with line 153-156.

(4) the authors compared the proposed AquaPile-YOLO method with some other methods, are they trained using the same dataset?

(5)Figure 15: in the caption, please explain what does the red box represent.

Response 4: 

(1)In response to your feedback, we have updated the manuscript to include the full name "CSP Bottleneck with 3 convolutions" for C3 at line 295.

(2)In response to your suggestion, we have removed the repeated full name of CBAM at line 298.

(3)In response to your feedback, we have revised and condensed the information at lines 374-376 in the updated draft. This revision aims to streamline the text and eliminate repetition, ensuring a more concise and clear presentation.

(4)We confirm that the AquaPile-YOLO method, as well as the other methods we compared it with, were all trained and validated using the same dataset. This dataset consists of 4,000 sonar images collected from field experiments, which encompass a wide range of underwater environments and target conditions. By using the same dataset for all methods, we ensure a fair and consistent evaluation of their performance. This approach allows us to accurately assess the relative strengths and improvements of the AquaPile-YOLO algorithm in terms of detection accuracy, robustness, and adaptability to various underwater scenarios.

(5)In response to your feedback, we have revised the caption to provide a more detailed explanation. The updated caption now reads: "Heatmap comparison of different algorithms for pile foundation detection by forward-looking sonar images. The heatmap illustrates the performance comparison of various detection algorithms, with the red box highlighting the area of interest where the pile foundation targets are detected. Within this box, the intensity of the color indicates the confidence level of the detection, with warmer tones (reds and yellows) signifying higher confidence in the presence of a target."

Author Response File: Author Response.docx

Reviewer 3 Report

Comments and Suggestions for Authors

See attached file. 

Comments for author File: Comments.pdf

Author Response

Comments 1: Suggest to add references to support the views and comments in Introduction Section, such as the first paragraph.

Response 1:  In response to your feedback, we have added relevant references to the first paragraph of the Introduction section, as seen in lines 27-35 of the revised draft. These references support the views and comments presented, providing a solid foundation for our research context.

Comments 2: Between 1st and 2nd paragraphs in the Section 1 of “Introduction”, suggest to add a summary of the primary studies concerning on the underwater target detection techniques based on the high-resolution sonar images. Then Summarize the main problems solved and to be solved. Suggest to include the content ‘aforementioned challenges by integrating multi-scale feature fusion and attention mechanisms’ you mentioned at Lines 38-39.

Response 2: In response to your feedback, we have added a new paragraph between the first and second paragraphs of Section 1 of the Introduction. This paragraph, now located at lines 43-54, provides an overview of the key studies in this field and highlights the primary challenges addressed and those that remain. Additionally, we have incorporated the content regarding the 'aforementioned challenges by integrating multi-scale feature fusion and attention mechanisms,' as you suggested.

Comments 3: The AquaPile-YOLO is proposed or improved in this study? If it is, suggest to improve the sentence at Lines 46-47. ‘the introduction of’ seems like the AquaPile-YOLO is from other studies.

Response 3: We appreciate your suggestion to improve the sentence at lines 46-47 to better reflect that AquaPile-YOLO is proposed in this study. In response to your feedback, we have revised the content at lines 117-120 in the updated draft. The revised sentence now clearly states that AquaPile-YOLO is proposed in this study, emphasizing our original contribution.

Comments 4: Why is the AquaPile-YOLO developed based on YOLOv5, not YOLOv6, v7, v8?

Response 4:  The decision to develop AquaPile-YOLO based on YOLOv5, rather than YOLOv6, YOLOv7, or YOLOv8, stems from a comprehensive evaluation of the algorithm's performance in terms of speed, stability, and accuracy—critical factors for real-time sonar image processing. YOLOv5 offers notable advantages in terms of computational efficiency and algorithmic stability, making it particularly suitable for real-time applications in underwater environments. Furthermore, our study incorporates advanced modules such as the C3N module from YOLOv7 into the YOLOv5 framework. Extensive experiments demonstrated that this hybrid approach effectively balances real-time detection speed, robustness, and accuracy, ultimately leading to the superior results presented in this paper.

Comments 5:Lines 66-68, I think the review paper is not a research paper, instead, the key studies in the review paper should be briefly introduced.

Response 5: In response to your comment, we have revised the content at lines 56-58 in the updated draft. This revision provides a concise introduction to the key studies discussed in the review paper, ensuring clarity and alignment with the expectations for a research paper.

Comments 6:Lines 68-69, modify the ‘B. R. Calder et al. and associates’. Does the ‘B. R. Calder et al.’ include the ‘associates’?

Response 6: In response to your feedback, we have removed the phrase 'and associates' from the citation, as it was redundant. The revised citation now appears at lines 58-59 in the updated draft. This change ensures clarity and precision in our references.

Comments 7: In Section 2.3, is it possible to summarize the advantages and disadvantages of traditional sonar image processing methods in Section 2.1, too?

Response 7:In response to your feedback, we have added a summary at lines 91-94 in the updated draft. This section now highlights how ongoing research integrates the strengths of traditional methods with advancements in deep learning to address existing challenges. The focus is on improving model generalization, efficiency, and robustness for sonar image processing applications.

Comments 8: Line 108, ‘Gaspar et al. and others’ is supposed to be ‘Gaspar et al.’?

Response 8: In response to your feedback, we have removed the phrase 'and others' from the citation. The revised citation now appears at line 97 in the updated draft. This change ensures clarity and precision in our references.

Comments 9: Line 110, is ‘PLUD’ an abbreviation?

Response 9: Yes, 'PLUD' is indeed an abbreviation. In response to your feedback, we have added the full expansion of the term as 'PLUD (Push the right Logit Up and the wrong logit Down)' at line 99 in the updated draft. This clarification should help readers better understand the concept we are referring to.

Comments 10: Suggest to specify the key problems resolved in this study for underwater pile foundation detection in Section 2.5 or somewhere.

Response 10:  In response to your feedback, we have added a detailed description of the key problems addressed in this study at lines 126-130. This section now clearly outlines the specific challenges we have tackled, such as improving detection accuracy, enhancing robustness in varying underwater conditions, and addressing the limitations of traditional sonar image processing methods.

Comments 11: Suggest to combine Figures 4, 5, 6, 7 into one figure. Suggest to add a colorbar for solar data in Figures 5 and 7.

Response 11: In response to your feedback, we have merged Figures 4 and 5 into a new Figure 4, and Figures 6 and 7 into a new Figure 5 in the updated draft. This consolidation should make the presentation more coherent and easier for readers to follow. Regarding the addition of a colorbar for solar data in Figures 5 and 7, we have considered this suggestion carefully. However, since the pseudo-color information in the sonar images is primarily used to enhance contrast and is not directly related to solar data, we have decided not to include a colorbar. The original sonar acquisition software does not provide a colorbar, as the pseudo-coloring is meant to improve visual interpretation rather than represent specific data values.

Comments 12: Line 202, what does the ‘yellow sonar images’ mean?

Response 12: The term "yellow sonar images" refers to sonar images that have been pseudo-colored using a yellow tone for enhanced contrast and visualization. This pseudo-coloring technique is applied during post-processing to improve the interpretability of grayscale sonar data, making features and target areas more distinguishable to the human eye.

Comments 13: Line 380, Check the title of Section 4

Response 13: I have revised it from "Experimental" to "Experiments".

Comments 14: Line 387, improve the sentence of ‘The experimental environment was a specific water area in a lake field test site’

Response 14: I have revised it to: "The experimental environment was a designated section of a lake field test site, characterized by water depths ranging from 2m to 20m and a substrate primarily composed of sand and gravel, providing a controlled yet representative setting for underwater sonar testing."(Line 376-379) This revision provides a more detailed and precise description of the experimental environment, highlighting the specific conditions that were conducive to our testing.

Comments 15: Line 569, check the ‘In the formula (XX)’

Response 15: I have corrected it from "In the formula (XX)" to "In the formula (13)".

Comments 16:Line 513, why do you choose the YOLOv3 in the comparison?

Response 16: We selected YOLOv3 primarily because its network architecture is relatively less complex compared to more recent versions like YOLOv5. This simplicity translates into faster processing speeds and greater stability, which are crucial factors for real-time underwater target detection using sonar imagery. In engineering applications where real-time detection is essential, the ability to process images quickly and reliably is paramount. YOLOv3's performance characteristics align well with these requirements, making it a suitable benchmark for evaluating the efficiency and effectiveness of our proposed AquaPile-YOLO algorithm in practical scenarios.

Comments 17:In figures 11-14, suggest to make the font size bigger and clearer. Suggest to combine them into one or two figures.

Response 17: In response to your feedback, we have made the following adjustments: Font Size and Clarity: We have increased the font size in the figures to ensure that the text is more legible and easier to read. This should enhance the overall clarity of the information presented. Combining Figures: Figures 11 and 12: We have merged the original Figures 11 and 12 into a new Figure 9. This consolidation results in a more succinct presentation of the experimental findings, enabling readers to quickly grasp the key results of our study. Figures 13 and 14: The original Figures 13 and 14 have been combined into a new Figure 10. This new figure focuses on presenting the critical experimental data, allowing for a more intuitive understanding of the study's outcomes.

Comments 18: Is there supposed to be a colorbar to show the different color meaning in Figure 15?

Response 18:  In response to your feedback, we have made the necessary adjustments. Specifically, we have added a "Heatmap Colorbar" to Figure 11 in the revised draft (which corresponds to the original Figure 15). This colorbar provides a clear indication of the different color meanings, enhancing the interpretability of the heatmap. Additionally, we have included a detailed annotation for the figure to further clarify the significance of the color variations.

Comments 19: I think Lines 594-610 belong to the Section of ‘Discussion’, and in ‘Conclusions’, these can be briefly summarized.

Response 19: In response to your feedback, we have made the following adjustments: Discussion : We have revised the content to focus more on the detailed analysis and interpretation of our findings, as well as the implications of these results within the broader context of underwater pile foundation detection. This section now provides a comprehensive discussion of the key issues and challenges addressed in our study. Conclusions: We have significantly condensed the 'Conclusions' section to provide a brief summary of the main findings and contributions of our research. This revised section highlights the key outcomes and their significance without delving into extensive detail, as suggested.

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

Thanks the authors for revising the manuscript and addressing my comments. The manuscript has certainly improved. The introduction section still needs some revision:

1. Line 38-41: as you have merged the previous two sections into one, you don’t need these sentences here to introduce what this section is about.

2. line 42-52 doesn’t make any sense. Firstly, it’s a bit repeated as the information provided in lines 53-76. Secondly, the objective of this study is included here, which is also repeated with lines 99-124. I would suggest to remove these lines.

3. you don’t need to list the contributions (line110-119) and the innovations (lines 124-140) in introduction section. You can merge these information with the first paragraph in Discussion. Normally, after presenting the existing studies and their drawbacks, in the last paragraph of introduction you need to present your research objectives.

4. there are some format problems for the reference, for example, line 56, “B.R. Calder et al.”,  I think this should be “Calder et al.”. Please double-check.

5. I strongly suggest the author to ask the co-authors, especially the corresponding author, to check and revise the manuscript carefully before it gets published.

Author Response

Comments 1:  Line 38-41: as you have merged the previous two sections into one, you don’t need these sentences here to introduce what this section is about.

Response 1: Following your suggestion, we have removed lines 38-41, which were introductory sentences for the merged section. The revised introduction has been streamlined to ensure a smoother flow and avoid unnecessary redundancy.

Comments 2: line 42-52 doesn’t make any sense. Firstly, it’s a bit repeated as the information provided in lines 53-76. Secondly, the objective of this study is included here, which is also repeated with lines 99-124. I would suggest to remove these lines.

Response 2: In response to your suggestion, we have carefully revised the manuscript to eliminate redundancy and streamline the content. Specifically, we removed the repetitive information in lines 42-52, as it overlapped with content in lines 53-76 and lines 99-124. The consolidated version now presents the study's objectives and background information in a more concise and coherent manner.

Comments 3: you don’t need to list the contributions (line110-119) and the innovations (lines 124-140) in introduction section. You can merge these information with the first paragraph in Discussion. Normally, after presenting the existing studies and their drawbacks, in the last paragraph of introduction you need to present your research objectives.

Response 3: Based on your suggestions, we have made the following revisions: 1.Reorganization of Contributions and Innovations: The content previously listed in lines 110-119 and lines 124-140 of the introduction has been removed and merged with the first paragraph of Section 4.3 (Discussion), as seen in lines 555-593 of the revised manuscript. This reorganization ensures that the introduction remains concise and focused while appropriately detailing the contributions and innovations in the discussion section. 2.Revised Research Objectives: The final paragraph of the introduction has been rewritten to clearly present the research objectives, following the presentation of existing studies and their limitations. This revision highlights the study’s goals and provides a logical transition to the subsequent sections.

Comments 4:  there are some format problems for the reference, for example, line 56, “B.R. Calder et al.”,  I think this should be “Calder et al.”. Please double-check.

Response 4: Yes, changing the sentence to:"Subsequently, Calder et al. presented a novel concept for underwater identification of side-scan sonar images—a Bayesian approach to target detection."is more appropriate. This revision aligns with standard citation practices in academic writing, where only the first author's surname is used followed by "et al." Similarly, other references, such as "G. L. Foresti et al.," have been corrected to follow the same format, using only the first author’s surname followed by "et al." where appropriate. These revisions ensure consistency and alignment with academic citation standards.

Comments 5:I strongly suggest the author to ask the co-authors, especially the corresponding author, to check and revise the manuscript carefully before it gets published.

Response 5: We have invited all co-authors, including the corresponding author, to carefully review and revise the manuscript to ensure its quality. Despite the time constraints, we are committed to thoroughly addressing any remaining issues in the paper. In addition to implementing the revisions based on your valuable feedback, we have made further improvements to enhance the manuscript. Specifically: 1.Revised Figures: We have restructured and resized all figures in the manuscript to maximize clarity and highlight the most critical information effectively. 2.Introduction Updates: Several sentences in the introduction have been refined for better readability and accuracy, including those in lines 31-36 of the revised manuscript. 3.Overall Quality Improvements: We have conducted additional checks to identify and address potential issues throughout the manuscript.

Author Response File: Author Response.docx

Reviewer 3 Report

Comments and Suggestions for Authors

Thanks for authors' efforts, please make all figures in the manuscript clearly to read.

Author Response

Comments:

Thanks for authors' efforts, please make all figures in the manuscript clearly to read.

Response :

We have restructured and resized all figures in the manuscript to maximize clarity and highlight the most critical information effectively.