Convolutional Neural Network-Based Method for Agriculture Plot Segmentation in Remote Sensing Images

Round 1

Reviewer 1 Report (Previous Reviewer 2)

Comments and Suggestions for Authors

1. Enhancement of Graphic Quality:

Upon review of the revised figures, I observed that the graphic resolution remains suboptimal, which could impede proper interpretation and analysis. High-resolution images are critical for a clear understanding of the data presented. I recommend that the authors employ a higher DPI setting (at least 300 DPI) for their figures to ensure that all details are legible and the images maintain their clarity in both online and print formats.

 

2. Addressing Previous Comments on Figures:

I note with concern that my previous comments concerning Figure 7 (now Figure 8 in the revised manuscript) have not been adequately addressed. It is imperative for the transparency of the revision process and the integrity of the scientific record that all reviewer comments are carefully considered and responded to in detail. I would urge the authors to provide a revised version of this figure that incorporates the suggested changes or a thorough explanation as to why they believe no modification is necessary.

 

3. Comprehensive Response to Reviewer Comments:

Lastly, I suggest that the authors provide a more exhaustive and itemized response to each of the review comments. This practice is essential not only for transparency but also for demonstrating how the reviewers' insights have been integrated into the revised manuscript. A point-by-point response can also facilitate a more efficient review process and allow for a clearer understanding of the revisions made.

 

In conclusion, while the authors have made certain revisions to the manuscript, addressing these specific concerns would greatly improve the quality and clarity of the research presented. I look forward to the authors’ response and an updated version of the manuscript for further consideration.

Author Response

Responses to your comments are included in the Word.

Author Response File: Author Response.docx

Reviewer 2 Report (Previous Reviewer 3)

Comments and Suggestions for Authors

Manuscript structure:

-          Page 4 Line 168: The reference section does not include citation [59].

-          Page 4 Line 171: Could you please provide the reference for 'Chen Tianhua and his colleagues' and include it in your citation?

-          Page 5 in the evaluation section: Please define the equation number for the sentence, such as '... in the result can be called Pixel Accuracy (PA) as shown in Equation (1),' and apply a similar format to all subsequent equations.

-          Authors may utilize the equation function in Microsoft Word to generate equations in the correct format.

-          Page 9 Line 389: In Figure 3, where is the area representing farmland indicated by the black color?

-          In my opinion, it may be appropriate to change the title of the section from 'Data Enhancement' to 'Data Augmentation,' as the content of this section primarily pertains to the process of data augmentation.

-          Page 10 Line 439: The authors should either use the equation function in Microsoft Word to represent (1 x e^(-4)) or replace it with the numerical value.

-          Please move the content from Lines 411 to 504 to the results section.

-          Page 13 Line 520: Please provide a citation for the reference from Kai Ming He's team.

-          Page 16 section 4.3:  What is the purpose of this section? I'm having difficulty understanding the meaning of this paragraph. Could you please clarify the intended message?

-          In Figure 10, The image demonstrates limited accuracy improvement between improved TransUNet, Remove CBAM, and Remove Deformable ConvNets. Could you please provide further clarification on TransUNet's performance?

-          I recommend conducting an ablation experiment for your proposed network architecture, including the following configurations:

o   TransUNet

o   TransUNet with residual module improvement

o   TransUNet with jump connection improvement

o   Improved TransUNet (TransUNet with both residual module and jump connection improvements)

This approach aligns with the methods you introduced in the 'Improved TransUNet' section.

-          I'm having trouble understanding the definitions of (g), (h), and other variables that the authors introduced in the results section. Could you please provide clarification?

-          Table 5 should present and explain the significance of the ablation experiments.

-          The manuscript's organization needs improvement. The contents in both Section 3 and Section 4 results should be integrated into the results section.

-          The explanation in the result section lacks clarity. I recommend revising it carefully to enhance understanding.

-          The discussion section lacks depth and comprehensive analysis. I suggest revising it carefully, particularly by addressing key points to provide a more meaningful and insightful discussion.

-          I recommend a thorough revision of the manuscript, paying particular attention to the organization of content within each section, and ensuring it aligns with the structure and guidelines established by the journal.

 

-          Please check for grammatical errors or consider using an English proofreading service.

Comments on the Quality of English Language

Moderate editing of the English language is necessary, and it is advisable to consider using an English proofreading service.

Author Response

Responses to your comments are included in the Word.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report (Previous Reviewer 2)

Comments and Suggestions for Authors

1.Clarity and Detail in Responses to Reviewers: The current response to the reviewer's comments appears too brief and lacks clarity. It would be beneficial for the authors to provide a more detailed explanation of the changes made in response to the feedback. Specifically, the authors should clearly outline each modification and directly quote or describe the revised sections in their response. This approach will help ensure that the changes address the reviewers' concerns adequately and transparently.

 

2.Technical Issue with Figure 1: I encountered difficulties in viewing Figure 1. It is crucial to ensure that all figures are accessible and display correctly for all readers. I recommend checking the file format and compatibility with common PDF viewers. If necessary, consider providing the figure in an alternative format or adjusting the resolution or size to enhance compatibility.

 

3.Presence of Blanks in the Paper: There are noticeable blanks within the document that disrupt the flow of reading and may indicate missing content or formatting errors. It is important to review the manuscript thoroughly to identify and rectify these gaps. This could involve checking for inadvertent omissions or correcting formatting issues that may have led to these blanks.

 

4.Structuring the Introduction Section: In academic writing, it is customary for the last subsection of the introduction to outline the organization of the paper, while the second-to-last subsection typically highlights the paper's contributions. This structure aids in guiding the reader through the manuscript and immediately clarifies the paper's value and layout. I recommend reorganizing the introduction section to align with this common format, as it will enhance the paper's readability and academic rigor.

 

Author Response

Answers can been seen in the Word.

Author Response File: Author Response.docx

Reviewer 2 Report (Previous Reviewer 3)

Comments and Suggestions for Authors

Manuscript structure:

-          Authors may utilize the equation function in Microsoft Word to generate equations in the correct format. Examples show in attached file. 

 -          Page 9 Line 389: In Figure 3, where is the area representing farmland indicated by the black color?

-          Page 10 Line 439: The authors should either use the equation function in Microsoft Word to represent (1 x e^(-4)) or replace it with the numerical value. Examples show in attached file. 

-          Please move the content from Lines 411 to 504 to the results section.

-          Page 16 section 4.3:  What is the purpose of this section? I'm having difficulty understanding the meaning of this paragraph. Could you please clarify the intended message?

-          Figure 6 and 7 should appear and explain in result section

-          In Figure 10, The image demonstrates limited accuracy improvement between improved TransUNet, Remove CBAM, and Remove Deformable ConvNets. Could you please provide further clarification on TransUNet's performance?

-          I recommend conducting an ablation experiment for your proposed network architecture, including the following configurations:

o   TransUNet

o   TransUNet with residual module improvement

o   TransUNet with jump connection improvement

o   Improved TransUNet (TransUNet with both residual module and jump connection improvements)

This approach aligns with the methods you introduced in the 'Improved TransUNet' section.

-          Table 5 should present and explain the significance of the ablation experiments.

-          The explanation in the result section lacks clarity. I recommend revising it carefully to enhance understanding.

 

-          The discussion section lacks depth and comprehensive analysis. I suggest revising it carefully, particularly by addressing key points to provide a more meaningful and insightful discussion.

Comments for author File: Comments.pdf

Comments on the Quality of English Language

Moderate editing of the English language is necessary, and it is advisable to consider using an English proofreading service.

 

 

Author Response

Answers can been seen in the Word.

Author Response File: Author Response.docx

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This paper investigated the application of convolutional neural networks (CNNs) to plot segmentation in remote sensing images. The authors first constructed a dataset for this task, and then applied four typical CNN architectures, including UNet, SegNet, DeepLabV3, and TransUNet, to this dataset. The authors studied and compared the different behaviors of these networks, and highlighted TranUNet for the plot segmentation task. Then the authors proposed to refine the residual module and jump connection in the TransUNet, obtaining improved performance. The authors also developed an interface for remote sensing image plot segmentation.

 

 

The studied problem is important, and the authors have contributed to it. However, there are several issues with this paper:

 

1. The organization is weird. In specific, according to my experience, the "Related Work" section should review literature related to the authors' work. In contrast, in this manuscript, the authors mainly described how they processed the data.

 

2. The authors spent lots of space to detailedly describe the convergence behaviors of the compared CNNs, which is not necessary since Figure 4 provides enough information for the convergence.

 

3. In Line 248, it is misleading to say the number of total training images is 7366, since according to Table 2, 7366 is the number of total images, while the actual number of training images should be 2550+2586=5136.

 

4. The captions of the figures are too long. 

Comments on the Quality of English Language

The writing could be improved. For example, there exist typos, such as that "sown" in Line 148 should be "shown". 

Author Response

The reply is attached. Please refer to the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

1. Abstract

The abstract is structured effectively, outlining the problem, the methods, and the results. I have some suggestions to enhance its clarity and content.

 

Problem Definition: The issue described in the opening sentence could be better clarified for readers unfamiliar with the term "agricultural plot." A suggested rewording is: "Accurate delineation of individual agricultural plots, the foundational units for agriculture-based activities, is crucial for effective government oversight of agricultural productivity and land utilization."

 

Methodology: While the methods applied are mentioned, it would be beneficial to elaborate more on the adjustments and enhancements made to the TransUNet network. This would give readers a clearer insight into your unique contribution to this field.

 

Results: The results section should explicitly state whether the presented percentages signify an improvement over pre-existing methods or the original TransUNet network. If so, what is the degree of improvement? 

 

There is a need to elaborate certain terminology, like IOU, when introduced for the first time.

 

2. Organization and Contribution

The structure of the paper could be enhanced by including a section devoted to the paper's organization and contributions.

 

To improve logical flow, it is advisable to introduce the method before presenting the experiment and its results.

 

3. Paper Presentation

In the related works section, the authors are encouraged to provide a more comprehensive review of other relevant studies. It's crucial to highlight their limitations that your work aims to address, as this will better justify and provide context for your research.

 

4. Methodology

The methodological improvement appears to mainly involve the reduction of training epochs. This might not offer sufficient novelty to your work. I recommend investigating additional ways to refine or advance the methodology.

 

5. Simulation and Results

Figure 7 appears to contain numerous inconsistencies, suggesting that the simulation may not have been executed optimally. It would be beneficial to revise this and ensure that all simulations are performed with the highest possible accuracy and precision.

Author Response

The reply is attached. Please refer to the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

Comments and Suggestions for Authors

Reviewer’s comments

English proofreading is necessary.

It would be better if you changed your topic to "Convolutional Neural Network-Based Method for Agriculture Plot Segmentation in Remote Sensing Images," which could enhance the focus of your research.

The abstract should be revised for logical structure, conciseness, brevity, and improved clarity of understanding.

The introduction needs to encompass a literature review and relevant research. This review should clearly define the gaps present in those studies, establishing a clear link to your proposed objectives aimed at addressing these gaps.

The final paragraph of the Introduction should encapsulate the unique contributions and innovations brought about by this research.

What novel contributions or innovations does this research bring forth?

The 'Related Works' section should encompass a comprehensive review of relevant literature, theories, and research pertaining to the proposed methodology.

The sections on Data Labeling and Data Enhancement should be situated within the Experimental Data section.

It is apparent that the researcher aims to introduce an improved TransUNet Network. This enhancement should be elaborated upon in detail within the Materials and Methods section.

The Materials and Methods section should not display the results obtained from the research, such as those presented in Table 4 and their corresponding explanations. This content should instead be included within the Results section.

Section 3.5 constitutes the focal point of this research. It should encompass a clearly defined heading and provide detailed step-by-step explanations

What is the intended benefit of Section 3.6 that you aim to convey? Was it previously outlined in the objectives of this study?

The experimental design needs to be appropriately restructured, following the correct sequence of topics.

Conducting an ablation experiment on the enhanced TransUNet network is recommended.

Your results are incomplete and lack clarity in reflecting your findings.

Where is your discussion? It does not appear in this manuscript.

The conclusion is incomplete.

 

Please ensure that your manuscript follows the logical structure of scientific writing and avoids mixing content in your manuscript.

Comments on the Quality of English Language

I find your English difficult to understand. I recommend that you consider using English proofreading services before submitting the revised version.

Author Response

The reply is attached. Please refer to the attachment.