Robust Coffee Rust Detection Using UAV-Based Aerial RGB Imagery
Round 1
Reviewer 1 Report
The manuscript is devoted to the detection of coffee rust by the UAV image analysis method.
Questions and comments on the Manuscript:
1. How does the cultivation of coffee in El Salvador correlate with the conflict between Russia and Ukraine (line 34)?
2. The citation [9-15] should be described in more detail with citation of 1-3 references.
3. What is the scientific novelty and practical significance of the authors' research in comparison with the works of [26-28]?
4. Why is it necessary to remove exactly 40% of the upper part when segmenting images, although Figure 7 shows that a significant part of the coffee plant image is removed?
5. I do not see the need to include a complete set of Figures in Additional Materials if they are already in the main text of the Manuscript.
6. Paragraphs 1-3 of the Discussion section (lines 294-313) are more appropriate in the Introduction than in the Discussion. The Discussion section needs to be improved.
7. Conclusions should be more specific and confirmed by numerical data.
Author Response
Dear reviewer,
Thank you for your questions and suggestions. We are attaching the review document.
Best regards.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Agrienginnering
Robust Coffee Rust Detection using UAV-based Aerial RGB Imagery
This paper described a method for coffee rust detecting. The developed algorithm was based on a two-stage approach. In the first stage, images are processed using ImageJ software. In the second phase, the morphological filters and the Hough transform were used by Python for rust identification. I have the following suggestions.
1. Too many backgrounds were described in the introduction, and the current research status was insufficient.
2. The case of the callout letters was not uniform between Figure 1 and the corresponding paragraph.
3. Parts 2.1 and 2.2 were recommended to be grouped together.
4. Figure 2 is not clear.
5. In Figure 4, the key components was recommended to highlight
6. The two images in Figure 7 do not indicate which was a and which was b.
7. “The top 40% of the image is cropped, along with 20% from each side.” What is the basis for crop**? Is it based on algorithmic identification or manual observation?
8. “Red and orange colors in the image obtained from ImageJ.” How was it obtained? I don't see an explanation.
9. How to distinguish between ordinary breakage and rust? Whether there will be a misjudgment.
Moderate editing of English language required
Author Response
Dear reviewer,
Thank you for your questions and suggestions. We are attaching the review document.
Best regards.
Author Response File: Author Response.pdf
Reviewer 3 Report
In this study, the method developed using ImageJ and Python software proved to be effective in evaluating coffee leaf rust detection. The obtained results were promising, indicating the potential application of this method to the RoCoLe dataset using drones. The research holds significance, and I have the following questions:
-
1. The drone's flight time was chosen from 8 AM to 10 AM. Does the lighting factor have any influence on the image quality? Please explain this aspect in your text.
-
2. Are there instances where some leaves are erroneously cropped or misclassified when proportionally crop** drone images? How can the problem of detecting coffee rust disease in tree leaf shadows be addressed?
-
3. The detection results do not show the coffee rust grade in the drone images.
-
4. What is the sequence of morphological operations in the opening and closing operations?
-
5. There are numerous studies on using drones for monitoring crop diseases. What is the novelty of this paper? Is it the drone remote sensing technology for detecting coffee rust disease or the image processing techniques?
-
6. In both datasets, coffee rust recognition did not match other methods. How is the superiority of the proposed method demonstrated?
-
7. With a plantation area of 75 hectares and a sample data of 96 photos at a distance of 5 meters above the ground, the representativeness is questionable. The article describes drone flights in relation to slopes, which seems inconsistent with Figure 3's depiction.
English expression needs to be improved appropriately.
English proficiency needs to be enhanced appropriately.
Author Response
Dear reviewer,
Thank you for your questions and suggestions. We are attaching the review document.
Best regards.
Author Response File: Author Response.pdf
Round 2
Reviewer 3 Report
In this study, the effectiveness of a developed method using ImageJ and Python soft- ware for coffee leaf rust detection was evaluated. The results obtained were promising when applying the method to both the RoCoLe dataset and images acquired using a UAV. This research has a certain significance.After revisions, I have no problems with the article.
