Feasibility of Early Yield Prediction per Coffee Tree Based on Multispectral Aerial Imagery: Case of Arabica Coffee Crops in Cauca-Colombia
Round 1
Reviewer 1 Report
The title of the manuscript does not satisfy with publication standard that does not reflect the originality. The introduction of the manuscript does not satisfy with publication standard that does not reflect the originality. Please clarify the originality of this research in the introduction from previous research.
Author Response
We want to thank the editor and the anonymous reviewers for their constructive and helpful comments. We have revised the manuscript by taking each comment into account.
Please see the attachment.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Major
This study is considered to be a thesis related to coffee crop prediction using aerial satellite imagery.
The data collection and data processing process for conducting this research are faithfully described.
If the results and considerations in this paper are supplemented, I think that it will be a paper of academic value.
It is suggested to create Results and Discussion parts separately.
What has been written now is judged by the results-oriented content.
The discussion of the data preprocessing process and results in this paper needs to be dealt with in depth.
Minor
Introduction
The first paragraph of Introduction was written too long. The first paragraph should be subdivided and revised to clearly reveal the purpose.
The full name of an abbreviation is often duplicated. Please check and correct. (ex. L80 RF)
Author Response
Many thanks to your comments enabled us to enhance the quality of this paper significantly. Our answers to your comments are in red below in this document
Author Response File: Author Response.pdf
Reviewer 3 Report
The accurate prediction of coffee production is very important for the sustainable development of the coffee industry. The author proposes a method to predict coffee production through multispectral aerial images. However, compared with previous work, this paper is not innovative enough, and fails to effectively explain the mechanism of physical and spectral characteristics in yield. Moreover, after reading the full text, I found that the author had many deficiencies in the language expression, data acquisition and chart making in this article. To sum up, this makes me unable to recommend the publication of this article.
In order to improve the quality of authors' manuscripts, the following aspects deserve attention:
Line 17, the abbreviations in the context shall be consistent;
Lines 55-67, it is necessary to sort out and summarize according to the data sources, characteristic parameters or methods of production forecast, rather than simply stating the work of previous scholars. This suggestion is applicable to the introduction;
Line 85-90, this article needs to be polished by experts with English mother tongue background to improve the writing of this article. The proposal is applicable to the full text;
Line 111-114, this part is unclear.
Figure 1&2: The drawing is not standardized, and the key information such as longitude and latitude, scale, orientation and legend is missing;
Line 123, it is recommended to change the order of sections 2.2 and 2.3;
Line 128, shouldn't it be cloudless and sunny weather? I doubt the quality of the image data.
Figure 3, please name it comprehensively according to the content expressed in the figure, rather than only indicating the method used therein. The proposal is applicable to the full text;
Line 172&182, "watered method", should be expressed consistently in the context.
Line 213, how to ensure the production accuracy of DSM? Why 10pixels, not others?
Formula (3), LAI? Should be diameter2?
Line 222-227, this part is common sense, and it is recommended to delete.
Table 1, Why are these vegetation indexes selected?
Table 2. How to filter the most relevant variables? What about hMed?
Figure 15: The description of horizontal and vertical coordinates is missing and the text is too small, so it should be drawn according to the requirements of the journal. This recommendation is applicable to the full text.
Table 3, Why do you choose these regression models? The results show that the prediction performance of the first four models is highly similar. In addition, in addition to the R2 side, this paper should use such indicators as MSE and RMSE to indicate the prediction error.
Author Response
Many thanks to your comments enabled us to enhance the quality of this paper significantly. Our answers to your comments are in red below in this document.
Author Response File: Author Response.pdf
Reviewer 4 Report
The manuscript "Predictors of Coffee Crop Yield based on Multispectral Aerial Imagery" (remotesensing-1975564), demonstrated the potential use of UVA and a monitoring using by vegetation index to determinate a predicted yield in coffea crop productions.
The authors have done a large amount of work, employing various references and critical analysis based on a scientific method and structure. The introduction, M&M, the results and discussion topic is good and minor points its necessary by adjusting in English synthases. Tables and figures are good qualities, but not the legends. In adition, legend of the figures and tables not represented displayed. I think, Authors should reformulate all legends of tables and figures, in an appropriate way, indicating what it is about.
My conception, the manuscript is suitable for publication, after few corrections. The structure is adequate, and the information is new and of great significance for comprehension.
In addition, I would just like to ask the authors a few questions, on small points that I was confused or that were not written in the manuscript.
Best regards
Points:
#01: There is a scope for improvement in the introduction section: a) additional emphasis on the significance of the study, b) scientific contribution of the paper in integration between other plants; Maybe a one paragraphs with potential economic by applied Multispectral Aerial Imagery to coffee in world, can be interestingly to Remote Sensing readers;
#02: How this methodology can be extended for other plants: potential challenges, advantages? Do you have any competitive disadvantages Multispectral vs Hyperspectral analysis in Coffea, based (economic, competition between species, contamination, changes in phenotypic plasticity in coffea)? Added in your results and discussion topic, please.
#03 Why, authors did not perform a principal component analysis or estimate the prediction by PLSR methods, for example? Why have not other statistical tools been tested, to improve the value of R2?
#04 why analyse only NDVI and CRI? Are other vegetation indices not more robust in the classification? Table 1, are not discuted in manuscript.
-Conclusion: What is future perspectives to use your Models (55 to 56%) in farms or industry to classified Arabica coffee crop?
English grammar syntax and sentences were necessary corrections.
Minor point:
-Check all scientific names;
-Warning, r2 or R2, Please superscript;
Figure 3 Segmentation? What is this? Please, check all legends in manuscript. It’s a not good!!!
Best regards,
Author Response
We want to thank the editor and the anonymous reviewers for their constructive and helpful comments. We have revised the manuscript by taking each comment into account.
Please see the attachment
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
It is ready for publication.
Reviewer 2 Report
Thank you for all the revisions which considerately improved your paper.
Reviewer 3 Report
The author proposes a method to predict coffee production through multispectral aerial images. However, compared with previous work, this paper is not innovative enough, this makes me unable to recommend the publication of this article.
