Detection of Tagosodes orizicolus in Aerial Images of Rice Crops Using Machine Learning

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Tagosodes orizicolus Muir is one of the most prevalent and destructive pests in rice crops, particularly in the tropical and subtropical regions of the Americas. Direct damage occurs through feeding and oviposition in the mesophyll and phloem. This study employs RGB images and machine learning techniques to detect areas infested by Tagosodes orizicolus in "Tinajones" rice crops during the flowering stage. The study has a certain degree of innovation. But I found some problems about it. Here is my comments.

1 The part of introduction should be improved. I haven't found any more relevant latest literature. Please introduce it.

2 The structure should be improved, which are introduction, materials and methods, results and discussion, conclusion.

2 In the part of related works, what is the difference of the related studies? Please explain it.

3 There are some English expression in the manuscript. It should be checked.

4 The part of abstract should be revised. I do not find the key results of the study. What is the scope of the study?

5 In the part of mateirals and methods, it should be introduced that the process of the experiment in detail, such as the time.

6 What is the meaning of fig.3? Please explain it.

7 The images are grouped into three classes. What is the standards?

8 There are mistakes in Eq(5).

9 Line213-227,The results of Fig4 should be explained.

10 Line 249, what is the meaning of F1-score of 0.963?

11 Line258-282, what is the results? I do not find.

12 The part of conclusion should be condensed, it is too long.

Comments on the Quality of English Language

The English should be polished by native speakers.

Author Response

Comments 1: The part of introduction should be improved. I haven't found any more relevant latest literature. Please introduce it.

Response 1: Thank you for pointing this out. The introduction has been improved by incorporating relevant and updated literature, which is presented in Table 1.

Comments 2: The structure should be improved, which are introduction, materials and methods, results and discussion, conclusion.

Response 2: The organization of the manuscript has been optimized to ensure it follows the standard structure of introduction, materials and methods, results and discussion, and conclusion.

Comments 3: In the part of related works, what is the difference of the related studies? Please explain it.

Response 3: A paragraph (lines 98-105) has been added to highlight the key differences between previous studies. Additionally, the discussion section elaborates on the specific distinctions of this study compared to prior research.

Comments 4: There are some English expression in the manuscript. It should be checked.

Response 4: The English expressions have been reviewed and corrected to enhance coherence and accuracy in the manuscript.

Comments 5: The part of abstract should be revised. I do not find the key results of the study. What is the scope of the study?

Response 5: The abstract has been updated to include the study's scope and the main results obtained.

Comments 6: In the part of materials and methods, it should be introduced that the process of the experiment in detail, such as the time.

Response 6: The experimental process has been detailed, including the date and time of image acquisition used in this study (see Table 2).

Comments 7: What is the meaning of fig.3? Please explain it.

Response 7: The description of Figure 3 (now figure 4) has been expanded in the Data Processing section, providing greater clarity regarding its significance.

Comments 8: The images are grouped into three classes. What is the standards?

Response 8: The classification of images into three categories has been detailed within the Data Processing section, explaining the criteria used.

Comments 9: There are mistakes in Eq(5).

Response 9: The error in equation (5) has been corrected.

Comments 10: Line213-227,The results of Fig4 should be explained.

Response 10: A more detailed description of the results represented in Figure 4 (now figure 5) has been added, and its title has been modified for better clarity.

Comments 11: Line 249, what is the meaning of F1-score of 0.963?

Response 11: The explanation of the F1-score has been improved in the corresponding section, providing more context regarding its significance in the study.
Comments 12: Line258-282, what is the results? I do not find.

Response 12: Additional content has been included in Section 3.2 (Application, lines 304-306), emphasizing the web application as a key outcome of the research and its practical relevance.

Comments 13: The part of conclusion should be condensed, it is too long.

Response 13: The conclusion has been revised, reducing its length while maintaining key information, in accordance with the reviewer's recommendation.

 

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Title: the title needs for specifying RGB images, name of two machine learning models plus 

web application developed. Using the rules of writing scientific name of sogata in all the manuscript such as in line 297

Authors' affiliation need to write with English language.

Abstract: abstract needs to add numerical results of two models for two sites (5 indicators P, Acc, R, S and FI).

Line 10 UAV write the full expression in the first one of the abbreviation (as in lines 51 and 111), apply to all manuscripts and if possible making a list of abbreviations.

Objectives of the work need specified and including the types of images, the name of two models or techniques and creation or developing of web application.

If possible adding some photos about sogata, (fig 3c doesn't enough) 

Define the TP, TN, FP and FN in this study to apply in the equation (2) of Acc.

Line 149 the summation of 75, 15 and 15 equal 105% from 500 images?

The results need to add more figs of all five indicators (P, R, S and FI) for two sites of study.

FI in paragraphs 239-250 is different with table 3 (please revisited most values) and this table 

need to more items to show these indicators for two sites especially environmental conditions were different (humidity, T and wind speed) according to table 2. How the value of precision of two models is equals (98.21) in table 3?

Add the chart of structure or architecture for two models.

Add the steps for processing to models.

Create flowchart and framework for this study.

Developed web application need to show and add the user interface and if possible add its link 

in the abstract and in the text line 259.

I recommended if possible in the end of this work, uploading this data on dataset websites for useful of scientific and academic community. 

With my best wishes,,

Author Response

Comments 1: Title: the title needs for specifying RGB images, name of two machine learning models plus web application developed.

Response 1: We appreciate the suggestion. We will wait for the editor's recommendations, and if necessary, we will make the adjustment accordingly.

Comments 2: Using the rules of writing scientific name of sogata in all the manuscript such as in line 297.

Response 2: The manuscript has been revised to replace "Sogata" with Tagosodes orizicolus, following the proper rules for writing scientific names throughout the document.

Comments 3: Authors' affiliation need to write with English language.

Response 3: The authors' affiliation has been corrected and is now written in English as: School of Systems Engineering, Señor de Sipán University, Chiclayo 14000, Peru.

Comments 4: Abstract: abstract needs to add numerical results of two models for two sites (5 indicators P, Acc, R, S and FI).

Response 4: Metrics related to model accuracy have been added, and we have highlighted that the results obtained are similar for both evaluated models.

Comments 5: Line 10 UAV write the full expression in the first one of the abbreviation (as in lines 51 and 111), apply to all manuscripts and if possible making a list of abbreviations.

Response 5: The use of the abbreviation "UAV" has been corrected throughout the manuscript, ensuring that its first occurrence includes the full term.

Comments 6: Objectives of the work need specified and including the types of images, the name of two models or techniques and creation or developing of web application.

Response 6: The introduction (lines 64-79) has been revised to clearly specify the study objectives, including the types of images used, the models applied, and the development of the web application.

Comments 7: If possible adding some photos about sogata, (fig 3c doesn't enough).

Response 7: More representative cropped images have been added. Since the study focuses on detecting affected areas using UAV images, the photographs display a broader area rather than individual plant details.

Comments 8: Define the TP, TN, FP and FN in this study to apply in the equation (2) of Acc.

Response 8: A more precise description of TP, TN, FP, and FN has been included (lines 205-211), specifying their application in Equation (2) for calculating accuracy (Acc).

Comments 9: Line 149 the summation of 75, 15 and 15 equal 105% from 500 images?

Response 9: Thank you for the observation. This was a typographical error; the correct distribution is 70%, 15%, and 15%, which has now been corrected in the text.

Comments 10: The results need to add more figs of all five indicators (P, R, S and FI) for two sites of study.

Response 10: The values in Table 3 have been updated, ensuring that each result is presented with three decimal places.

Comments 11: FI in paragraphs 239-250 is different with table 3 (please revisited most values) and this table 

Response 11: 
The difference between the values in the description and those in the table is because the former corresponds to results obtained for each class based on the applied model, while the table presents the average values for all classes for each algorithm.
The values in Table 3 have been reviewed and corrected for consistency.

Comments 12: need to more items to show these indicators for two sites especially environmental conditions were different (humidity, T and wind speed) according to table 2. How the value of precision of two models is equals (98.21) in table 3? 

Response 12: The values in Table 3 have been updated. The results are very similar because both models are trained with the same images.

Comments 13: Create flowchart and framework for this study.

Response 13: Figure 1 has been added, illustrating the process flow developed for detecting infested areas in this study.

Comments 14: Developed web application need to show and add the user interface and if possible add its link in the abstract and in the text line 259.

Response 14: The user interface has been updated in Figures 7 and 8, displaying the web interface within the browser.
A reference link to the web application has been added in the abstract and in line 259 of the text to facilitate access.

Comments 15: I recommended if possible in the end of this work, uploading this data on dataset websites for useful of scientific and academic community.

Response 15: The authors will make the data available to those who request it via email.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript has been modified according to the comments. I agree it can be accepted.

Author Response

We sincerely appreciate your feedback and the opportunity to improve our manuscript. Thank you for your review and acceptance.

Reviewer 2 Report

Comments and Suggestions for Authors

There are many comments don't response for all of it, which it's important to show. The authors can revise my previous comments and response for it (especially data and results of two sites)  

Author Response

Comments 1: Title: the title needs for specifying RGB images, name of two machine learning models plus web application developed.

Response 1: We appreciate the suggestion. We will wait for the editor's recommendations, and if necessary, we will make the adjustment accordingly.

Comments 2: Web application developed. Using the rules of writing scientific name of sogata in all the manuscript such as in line 297

Response 2: The manuscript has been revised to replace "Sogata" with Tagosodes orizicolus, following the proper rules for writing scientific names throughout the document.

Comments 3: Authors' affiliation need to write with English language.

Response 3: The authors' affiliation has been corrected and is now written in English as: School of Systems Engineering, Señor de Sipán University, Chiclayo 14000, Peru.

Comments 4: Abstract: abstract needs to add numerical results of two models for two sites (5 indicators P, Acc, R, S and FI).

Response 4: In the abstract we have considered only adding the precision of all the metrics, in the results section we can see all the values.

Comments 5: Line 10 UAV write the full expression in the first one of the abbreviation (as in lines 51 and 111), apply to all manuscripts and if possible making a list of abbreviations.

Response 5: The use of the abbreviation "UAV" has been corrected throughout the manuscript, ensuring that its first occurrence includes the full term.

Comments 6: Objectives of the work need specified and including the types of images, the name of two models or techniques and creation or developing of web application.

Response 6: The introduction (lines 64-79) has been revised to clearly specify the study objectives, including the types of images used, the models applied, and the development of the web application.

Comments 7: If possible adding some photos about sogata, (fig 3c doesn't enough).

Response 7: More representative cropped images have been added. Since the study focuses on detecting affected areas using UAV images, the photographs display a broader area rather than individual plant details.

Comments 8: Define the TP, TN, FP and FN in this study to apply in the equation (2) of Acc.

Response 8: A more precise description of TP, TN, FP, and FN has been included (lines 205-211), specifying their application in Equation (2) for calculating accuracy (Acc).

Comments 9: Line 149 the summation of 75, 15 and 15 equal 105% from 500 images?

Response 9: Thank you for the observation. This was a typographical error; the correct distribution is 70%, 15%, and 15%, which has now been corrected in the text.

Comments 10: The results need to add more figs of all five indicators (P, R, S and FI) for two sites of study.

Response 10: The values in Table 3 have been updated, ensuring that each result is presented with three decimal places.

Comments 11: FI in paragraphs 239-250 is different with table 3 (please revisited most values) and this table 

Response 11: 
The difference between the values in the description and those in the table is because the former corresponds to results obtained for each class based on the applied model, while the table presents the average values for all classes for each algorithm.
The values in Table 3 have been reviewed and corrected for consistency.

Comments 12: need to more items to show these indicators for two sites especially environmental conditions were different (humidity, T and wind speed) according to table 2. How the value of precision of two models is equals (98.21) in table 3? 

Response 12: The values in Table 3 have been updated. The results are very similar because both models are trained with the same images.

Comments 13: Add the chart of structure or architecture for two models.

Response 13: Figure 6 and 7 with Convolutional Neural Network Architecture VG16 and RestNet50 have been added.

Comments 14: Add the steps for processing to models.

Response 14: The steps for each architecture have been added in section 2.4.

Comments 15: Create flowchart and framework for this study.

Response 15: Figure 1 has been added, illustrating the process flow developed for detecting infested areas in this study.

Comments 16: Developed web application need to show and add the user interface and if possible add its link in the abstract and in the text line 259.

Response 16: The user interface has been updated in Figures 9 and 10, displaying the web interface within the browser.
A reference link to the web application has been added in the abstract and in line 259 of the text to facilitate access.

Comments 17: I recommended if possible in the end of this work, uploading this data on dataset websites for useful of scientific and academic community.

Response 17: The authors will make the data available to those who request it via email.

We appreciate once again your valuable contribution and hope to have your approval.

Author Response File: Author Response.pdf