Next Article in Journal
Collision-Free 4D Dynamic Path Planning for Multiple UAVs Based on Dynamic Priority RRT* and Artificial Potential Field
Previous Article in Journal
Pollution from Highways Detection Using Winter UAV Data
 
 
Article
Peer-Review Record

Animal Detection and Counting from UAV Images Using Convolutional Neural Networks

by Kristina Rančić 1, Boško Blagojević 2, Atila Bezdan 2, Bojana Ivošević 3, Bojan Tubić 4, Milica Vranešević 2, Branislav Pejak 3, Vladimir Crnojević 3 and Oskar Marko 3,*
Reviewer 1:
Reviewer 2: Anonymous
Reviewer 3: Anonymous
Submission received: 30 December 2022 / Revised: 27 February 2023 / Accepted: 1 March 2023 / Published: 6 March 2023

Round 1

Reviewer 1 Report

1.       What is this [?] in line no 76?

2.       Relate the scope of research with the importance of using AI models in the introduction section.

3.       Conclude the literature review with the research gap which leads you to do this research.

4.       What is 1 represent in line 228? Is it figure no or what? Make it clear.

5.       What is the size of the data set after annotation?

6.       Add annotated images with the original image of any 1 image from the data set.

7.       Add the table of model summary of each model.

8.       Why didn’t you use the latest versions of YOLO?

9.       Graphs in figure 6 and figure 7 are not proper. The peaks are not shown properly. If the graphs are correct. Kindly elaborate on them.

10.   Figure 8, Figure 9, Figure 19, and figure 20 are blurred which shows the data set is not having clear images.

11.   Graphs in Figures 12, 13, 15, and 16 cannot be visualized properly. Readings are blurred.

12.   The first image of figure 17 is labeled improperly as it considers 2 deer as 1 and covers the shadow of deer too.

Summary: Need to map this research with SDG’s of UNO.

Author Response

Dear reviewer,

Please find the attachment that contains answers to your questions. 

Best regards.

Author Response File: Author Response.pdf

Reviewer 2 Report

Please check the attachment.

Comments for author File: Comments.pdf

Author Response

Dear reviewer,

Please find the attachment that contains answers to your questions. 

Best regards.

Author Response File: Author Response.pdf

Reviewer 3 Report

In this work, authors have used convolutional neural networks for the detection and counting of animals in UAV images. Authors used different object detection architectures for the detection of deer in UAV images. Finally, a comparative analysis of all the object detection algorithms was performed for deer detection.

·         Page number 2, line number 62-63, rewrite the sentence.

·         Page number 2, line number 76, citation missing.

·         Page number 3, line number 95, citation missing.

·         In the introduction section, authors mentioned about the previous studies in this area but not included anything about the limitations/drawbacks of those studies.

·         Authors are suggested to explicitly mention the contributions made by them in this study.

·         It is suggested to mention how their work is different from the existing studies.

·         It is suggested to add a table of specifications of UAVs used in this study.

·         Authors are suggested to add flight planning parameters for all the flights conducted for this study.

·         It is suggested to add an overall methodology flowchart in the manuscript.

·         Authors have considered only 4 versions of YOLO for object detection. Why not considered the latest version of YOLO family such as YOLOv5, v6, v7?

·         Page number 18, line number 472, rewrite the sentence.

·         Page number 18, line number 517, citation missing.

·         Figure number 19 is not cited anywhere in the text of this manuscript.

·         The accuracy obtained for deer detection in this study is still low, authors are suggested to add more training data and also a variety of training images can be included which will be helpful in improving the accuracy of the model.

·         How the presented work is different from the work presented in [9].

Author Response

Dear reviewer,

Please find the attachment that contains answers to your questions. 

Best regards.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

It is advised to the author make a compliance reply and highlight all the comments given by me. Moreover, include new suggestions with my previous comments and answer the following carefully; 

0. Why the authors use YOLO old version 

1. References have still some issues check it carefully.

2. Delete archiv based references and conference its not suited 

 

Author Response

Dear reviewer, 

Please find the attachment as requested for changing the manuscript according to your suggestions. All changes in the manuscript are highlighted.

Best regards,
Oskar

Author Response File: Author Response.pdf

Reviewer 2 Report

The author has done a good job with the revisions. In addition, I have a few remaining further questions/remarks.

 

1.       Reference 23 was an error.

2.       In Introduction: This manuscript should further add some articles about YOLO models and will be of interest to many, such as:

(1)  Livestock detection in aerial images using a fully convolutional network. https://doi.org/10.1007/s41095-019-0132-5.

(2)  Automatic recognition of pavement cracks from combined GPR B-scan and C-scan images using multiscale feature fusion deep neural networks, https://doi.org/10.1016/j.autcon.2022.104698.

3.       Please check the picture quality again to meet the publication requirements.

Author Response

Dear reviewer,

We have taken into account your suggestions and please find the attached document with answers. 

Best regards,
Oskar

Author Response File: Author Response.pdf

Reviewer 3 Report

Authors have addressed all the comments.

Author Response

Dear reviewer,

We gratefully thank You for Your outstanding comments and suggestions about our work and manuscript. Thank you for taking the time to review our work. 

Best regards,
Oskar

Back to TopTop