Monitoring System for Leucoptera malifoliella (O. Costa, 1836) and Its Damage Based on Artificial Neural Networks

Round 1

Reviewer 1 Report

These are my main comments on the manuscript (agriculture-2100630) entitled “Development of a monitoring system for Leucoptera malifoliella (O. Costa, 1836) and its damage based on artificial neural networks”. Following substantial revisions should be incorporated in the manuscript prior to acceptance.

1. I have concerns about the manuscript sections that I believe need to be addressed in order to improve its clarity.

2. A hypothesis for this work is needed.

3. In methods, authors should explain the statistical methods used in each experiment.

4. Results and discussion should be divided in two sections, without this the manuscript cannot published.

5. Other revisions could be checked in PDF attached.

Comments for author File: Comments.pdf

Author Response

Dear Reviewer 1,

Please see the attachment.

Sincerely,

Dana Čirjak, M.sc.

University of Zagreb, Faculty of Agriculture

Department for Agricultural Zoology

25 Svetosimunska Street

Zagreb 10 000

Croatia

Author Response File: Author Response.docx

Reviewer 2 Report

The authors developed an ANN-based approach for pest control and monitoring. Some major revisions are required.

1) The introduction is well introduced. However, previous work on pest monitoring should be discussed extensively. The following study can be used as an example.

https://doi.org/10.1007/s12161-022-02251-0

2) "e" images are not explained in Figure 3 caption.

3) More citations are needed for section 2.3.

4) Equations are required for the metrics in Table 2 and Table 3. The following article can be used.

https://doi.org/10.1016/j.bspc.2021.102716

5) Figure 7 and Figure 8 should be clearer. ROI values are not read. Also, Figure 8 contains many detection frames, is that correct?

6) A figure showing the methodology steps of the proposed method is required.

7) Why did ANN come to the fore in the title and text while deep learning was used in the study?

8) Why is Section 3.3 explained in the results? Isn't that material information?

9) Conclusion section is insufficient.

10) The contributions of the study should be explained.

Author Response

Dear Reviewer 2,

Please see the attachment.

Sincerely,

Dana Čirjak, M.sc.

University of Zagreb, Faculty of Agriculture

Department for Agricultural Zoology

25 Svetosimunska Street

Zagreb 10 000

Croatia

Author Response File: Author Response.docx

Reviewer 3 Report

The paper explains about the concepts embodied below

1.     The aim of this study is to develop two models using artificial neural networks (ANNs) and two monitoring devices with cameras for early detection of L. malifoliella (Pest Monitoring Device) and its mines on apple leaves (Vegetation Monitoring Device) with emphasis on premature defoliation and as an implementation in the domain of automatic pest monitoring systems.

2. The collection of RGB images is done in order to learn (or train) two ANNs and to identify two main classes: 1. the pest, L. malifoliella; and 2. the damage caused by this pest (leaf mines).

3. For establishing a model for detecting L. malifoliella adults in PMD, three classes were defined and annotated: MINER, INSECT and OTHER.

 4. For establishing a model for detecting mines on apple leaves, two classes: MINES i.e., damage caused by L. malifoliella, and OTHER i.e., other objects.

I want to express the paper better in the following lines

1.       The paper could have been explained more clearly.

2.       How 144,000 images for training and 144,000 images for testing is obtained shall be explained further thoroughly.

3.       It is claimed that images are taken manually and on the contrary, it is also mentioned that camera is operated automatically.

4.       It could not be ascertained clearly whether the two ANN models are EfficientDet Object detection model itself or any other ANN model is also used.

5.       In my view in addition to the count, the expert entomologist report could be used as area of mine’s ground truth, and area of ANN analysis of mines could have been compared.

6.      The paper is definitely a major improvement in the evolution of Automatic Pest Management system.

 

Author Response

Dear Reviewer 3,

Please see the attachment.

Sincerely,

Dana Čirjak, M.sc.

University of Zagreb, Faculty of Agriculture

Department for Agricultural Zoology

25 Svetosimunska Street

Zagreb 10 000

Croatia

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

The authors have incorporated all suggestions and comments into the revised version, now the manuscript seems much clear. There is minor point to be corrected:

L.150: … study is divided…

L.152: … phase, data were processed, and the photos labeled…

Ls.153,155,170: Change is/are by was/were

L.171: The two classes were: 1. The…

Ls.277-415: in these paragraphs, sentences should be in the past form. Revise

Results: Again, check sentences in past form

 

 

Author Response

Dear Reviewer 1,

Please see the attachment.

Sincerely,

Dana Čirjak, M.sc.

University of Zagreb, Faculty of Agriculture

Department for Agricultural Zoology

25 Svetosimunska Street

Zagreb 10 000

Croatia

Author Response File: Author Response.docx

Reviewer 2 Report

The authors made the necessary corrections. But more work should be cited for the previous first comment. Current literature knowledge and comparison are important and necessary. The following study should be used.

A Novel Convolutional-Recurrent Hybrid Network for Sunn Pest–Damaged Wheat Grain Detection

Author Response

Dear Reviewer 2,

Please see the attachment.

 

Sincerely,

 

Dana Čirjak, M.sc.

University of Zagreb, Faculty of Agriculture

Department for Agricultural Zoology

25 Svetosimunska Street

Zagreb 10 000

Croatia

Author Response File: Author Response.docx