The High-Precision Detection Method for Insulators’ Self-Explosion Defect Based on the Unmanned Aerial Vehicle with Improved Lightweight ECA-YOLOX-Tiny Model

Round 1

Reviewer 1 Report

In general, I appreciate your work on this topic. The results are very clear. However, I have a few remarks about paper improvements, to have better "Scientific Soundness" and Quality of Presentation.

1. The main paper about this topic [1] clearly describes methodology and your paper claims that various versions of YOLO-X -tiny can serve for practical implementation in drone's payload. Still, it is not clearly justified what will be gained by YOLO X-Tiny in comparison to YOLO-X-X, YOLO-X-Nano, YOLO-R, etc. because future processors might easily cope with their' execution requirements. Having such paper, the future implementor can easily choose whether a more powerful processor is needed or not. 

Suggestion: add results from other more processing power requiring algorithms and prove that there is no significant differences for this application. This means just simple expansion of Table III.

2. Is there any way to process results that are not recognized by this methodology?  The idea that first comes to my mind is: by using results from Fig 12 and Fig 13 some decision logics that ambiguity cases should be recorded and sent for further inspection.

3. Example pictures of not correctly recognized cases should be displayed. Maybe these cases will be detected by using e.g. YOLO-X-X which will justify the need for better processors.

4. Based on 3., maybe some idea how to implement mentioned logic (or some more complex algorithm) I have skeched in 2  (I don't insist, but this will considerably rise the contribution of this paper)   

3. Yolo-X tiny is already described in references. I must admit that its description in the presented paper helped me to read, but I suggest putting it in the Appendix. Current paper structure can drag away the reader from the real results and contribution of this paper. This paper will look completely OK even if section II THE CONSTRUCTION OF ECA-YOLOX-TINY NETWORK MODEL is completely omitted. Still for reading easiness I recommend that it should be present, but in the appendix 

 

 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

My suggestions and concerns are as below:

 

1. All of the figures should be redrawn and prepared with high resolutions. Any figures which are given from a reference should be cited. 

 

2. A table including similar studies should be prepared and different features of the study should be compared accurately.

 

3. Introduction section can be detailed by reading and considering new publications, especially in Applied Sciences.

 

4. Accurately identify what is the main novelty of your study, especially with the below publication:

 

Qiu, Z.; Zhu, X.; Liao, C.; Shi, D.; Qu, W. Detection of Transmission Line Insulator Defects Based on an Improved Lightweight YOLOv4 Model. Appl. Sci. 2022, 12, 1207. https://doi.org/10.3390/app12031207

 

Some of the sentences in your paper as the same as the above paper. Let us know the differences and novelties.

 

Sincerely

 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

1.Figure 1 and Figure 10 poor in quality 

2. References section must be improved with recent related works published already.

3.What is inference you got from Figure 11.a and 11.b.

4.author contribution is not explained in details

5. comparison with existing system is missing , not given in detail

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

The new version is much improved. I have no further remarks. 

Reviewer 2 Report

This version is acceptable.