Analysis of Resampling Methods for the Red Edge Band of MSI/Sentinel-2A for Coffee Cultivation Monitoring

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Abstract:

1) The objective is not clear. It should be more specific.

1. Introduction

2) Remove Table 1 from the introduction and place it in the materials and methods section.

 

2. Materials and Methods

3) There are two Figure. Rename them according to the sequence in which they appear in the text.

4) Identify the location of the farms in the image with the title: “These are original B05 images: a) “Ouro Verde” Farm image; b) “Canto do Rio” Farm image.”

5) Indicate the formulas for the following spectral indices: NDRE (Normalized Difference Red Edge Index), the CCCI (Canopy Chlorophyll Content Index), and the IRECI (Inverted Red Edge Chlorophyll Index).

 

 5. Conclusions

6) In the conclusion section, there is a missing perspective related to future research work

7) Please write about the limitations of this work in details in conclusion section?

 

Author Response

Dear Reviewer,
We appreciate your time and effort in reviewing our manuscript. Below, we provide a detailed response to each of your comments and describe the corresponding revisions made:
Abstract:
1. The objective is not clear. It should be more specific.

*The objective has been rewritten to provide a clearer and more specific description of the study's purpose and scope.

Introduction:
2. Remove Table 1 from the introduction and place it in the materials and methods section.
*Table 1 has been moved to the Materials and Methodssection as suggested.

Materials and Methods:
3. There are two Figures. Rename them according to the sequence in which they appear in the text.
*The figures have been renamed in the correct sequential order.
4. Identify the location of the farms in the image with the title: “These are original B05 images: a) ‘Ouro Verde’ Farm image; b) ‘Canto do Rio’ Farm image.”
*The requested title has been added, explicitly identifying the farm locations in the images.
5. Indicate the formulas for the following spectral indices: NDRE
(Normalized Difference Red Edge Index), CCCI (Canopy Chlorophyll
Content Index), and IRECI (Inverted Red Edge Chlorophyll Index).
*The formulas for NDRE, CCCI, and IRECI have been included in the text.

Conclusions:
6. In the conclusion section, there is a missing perspective related to
future research work.
*In addition to exploring machine learning-based resampling techniques, we now suggest future studies assess how different
resampling methods perform under various environmental conditions and coffee production systems.
*This perspective aims to contribute to the refinement of remote sensing strategies for precision agriculture.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Summary:

The study evaluates different resampling methods (nearest neighbor, bilinear, cubic, cubic spline, and Lanczos) for the Red Edge band of Sentinel-2 in coffee cultivation monitoring. Using 500 random points from two farms in Brazil, it compares methods based on R², MAE, and RMSE, finding that cubic interpolation provides the best accuracy. Although the study is very effective and has practical implications, I would like to suggest a major revision at this stage. Please see my detailed comments below. If these comments can be addressed effectively, I can consider the manuscript for acceptance in the next round.

1. The selection of nearest neighbor, bilinear, cubic, cubic spline, and Lanczos methods needs effective justification. Why were these methods chosen? Were other interpolation techniques (e.g., Kriging, deep learning-based resampling) considered? You do not need to apply them at this stage. I am only interested in knowing if they were considered. If not, it can be highlighted in the limitations section in discussion somewhere
2. The methodology should specify whether the resampling was applied uniformly across both images (e.g., same kernel size, same interpolation window). Was any preprocessing (e.g., radiometric or atmospheric correction) performed before resampling?
3. You mentioned that B05 images were acquired from Sentinel-2 on two dates, but the processing level (e.g., L1C, L2A) is only briefly mentioned. Were images pre-processed (e.g., converted to surface reflectance), or were they used as is? If so, which corrections were applied? Or did you use an already processed product such SR
4. 500 random points were generated. Were these points distributed evenly across the study sites? Was any stratification or spatial uniformity considered? If yes, please provide more details
5. You used cross-validation with a 70/30 train-test split, but the paper does not specify whether the split was performed spatially or randomly. Was k-fold cross-validation used beyond the 70/30 split? If so, how many folds were considered? If not, it is fine, but explanation can improve.
6. You mentioned R and the terra package but does not provide the full computational setup. What hardware was used (e.g., CPU, GPU)? Were computations performed on a local machine or a cloud-based system?
7. As mentioned in the earlier comments, methodology does not acknowledge any limitations of the approach. Were there challenges in handling mixed pixels or artifacts introduced during resampling? How do these affect the interpretation of results? If not in methods, you can put some limitations of current research in discussion section

Author Response

Dear Reviewer,
We sincerely appreciate the detailed feedback and constructive suggestions from the reviewer. Below, we outline the revisions made in response to the comments:

1. Rationale for the choice of resampling methods:
* We have expanded the explanation in the Materials and Methodssection to justify the selection of the nearest neighbor, bilinear, cubic, cubic spline, and Lanczos methods. These methods were chosen based on their widespread application in remote sensing and their distinct interpolation characteristics, which allow for a comprehensive assessment of their impact on coffee crop monitoring.
*Additionally, we acknowledge that alternative techniques, such as Kriging and deep learning-based resampling, were not included in this study but are now mentioned as potential areas for future research.

 

2. Uniformity of resampling application:
* We clarified that resampling was applied uniformly across both images, ensuring consistency in kernel size and interpolation window across all methods.

3. Sentinel-2 B05 image preprocessing:
*The manuscript now explicitly states that no additionalradiometric or atmospheric corrections were needed for resampling, as Sentinel-2A (L2A) imagery already includes these corrections.

 4. Distribution of the 500 random points:
*We have added more details on how the random points weredistributed across the two farms, specifying whether spatial stratification or uniformity was considered in their placement.

 5. Training-test split methodology:
*The manuscript now clarifies whether the 70/30 training-test split was performed randomly or spatially. Additionally, we specify whether k-fold cross-validation was used.

 6. Computational setup:
*Information on the hardware used for the study has been included, specifying whether the computations were performed on a local machine or a cloud-based system.

 7. Limitations of the approach:

*We acknowledge that potential challenges related to mixed pixels and artifacts introduced during resampling were not discussed in this manuscript.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

Some changes are necessary to make the article publishable. The Introduction section should not contain tables 1 and 2...they should be moved to the next section perhaps. 

I ask the authors to solve a problem related to the presence along with the results of personal opinions moving them in Discussion section.

The Figure 5 and 6 should be enlarged to make the numbers more visible.

Comments for author File: Comments.pdf

Author Response

Dear Reviewer,
We sincerely appreciate the reviewer’s comments and suggestions, which have helped us improve the clarity and structure of the manuscript. Below are our responses to the requested revisions:

1. Relocation of Tables 1 and 2:
*As suggested, we have moved Tables 1 and 2 from the Introduction section to the Materials and Methods section to ensure proper structuring of the manuscript.

 2. Separation of personal opinions from results:
*We have revised the Results section to remove subjective statements and have relocated any interpretative or opinion-based comments to the Discussion section, ensuring that the results are presented objectively.

 3. Adjustment of Figure 5 and 6:
*Figures 5 and 6 have been enlarged to improve the readability of numerical values, ensuring that the information is more clearly visible to the reader.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The comments were addressed. Minor style issues need to be checked.

Author Response

Dear Reviewer,

We have reviewed the manuscript and made the necessary corrections. They are marked in red.

- Introduction: We reviewed the section to ensure that all relevant references were included and that the background information was sufficient.

- Research Design: Adjustments were made to improve clarity and coherence in the methodological framework of the study.

- Methods: The descriptions were refined for better accuracy and completeness, ensuring clarity in the approach used.

- Results: The presentation of the results was improved for better readability and consistency.

- Conclusions: We ensured that all conclusions were well supported by the findings of the study.

In addition, we reviewed the manuscript for minor stylistic issues and corrected them as necessary.

We appreciate your valuable contribution to improving the quality of our work.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

In the previous round of review, I asked the authors to solve a problem related to the presence along with the results, of personal opinions moving them in Discussion section. Many of these sentences, indicated in the specific comments, remained as they were. 
Therefore, I stand by my previous judgement: this paper needs major revision.

Comments for author File: Comments.pdf

Author Response

Dear Reviewer,

We appreciate your comments and suggestions for improving our manuscript. Below, we address each of the points and detail the corresponding revisions made to the text, highlighted in red.

Introduction: We ensured that the introduction provided sufficient context and included references relevant to the study.

Research Design: We considered the research design to be more appropriate after the modifications.

Methods Section: We revised and improved the description of the methods, clarifying aspects as necessary.

Presentation of Results: We refined the presentation of the results to increase clarity.

Conclusions: We ensured that the conclusions are supported by the results.

Specific Comments:

Line 30: The sentence "Studies on remote sensing (RS) in coffee production..." was revised for clarity.

Line 41: "Among these geotechnologies is orbital Remote Sensing (RS)..." was revised for accuracy.

Lines 47-49: The original references for the indexes mentioned have been included and their formulas have been removed as requested.

Lines 68-69: The phrase "RS of coffee farming" has been revised and refined.

Lines 76-77: The reference supporting the statement "Although the effects of these methods are well documented in the literature..." has been added.

Line 77: "RS of coffee" has been adjusted for clarity.

Line 83: The phrase "Other studies [30]... only one?" has been improved for consistency.

Line 90: Additional references have been inserted to support "... in RS studies...".

Line 141: The acronym "RS" has been checked and adjusted for consistency.

Lines 173, 190-191, 204, 207-208, and 223: These phrases have been removed from the article. We thought it best not to include them in the conclusion.

Line 232: The sentence "The mean (danger) value is nearly zero, suggesting the absence of..." has been removed.

Lines 291-299: The observations in this section have been removed from the original positioning and incorporated into the conclusion with appropriate adaptations.

References:

The references have been checked and formatted according to the required style guidelines.

We sincerely appreciate your insightful analysis, which contributed significantly to improving the manuscript. Please let us know if any additional revisions are needed.

Author Response File: Author Response.docx