Real-Time Object Detection Based on UAV Remote Sensing: A Systematic Literature Review

Round 1

Reviewer 1 Report

This paper systematically reviews UAV-based real-time object detection, addressing the limited focus on onboard implementation in prior research. It offers a comprehensive hardware and software framework, enhancing its value as a reference for UAV real-time data processing. The paper also highlights the necessity of real-time detection in various application scenarios and investigates the impact of sensors, computing platforms, algorithms, and paradigms on accuracy and speed in UAV real-time object detection.

Overall, the paper is well-written; however, I recommend the authors conduct a final grammar check. Additionally, please consider the following suggestions for improvement:

1. Correct the terminology in Figure 5 to "Sankey diagram" rather than "Traffic diagram" to align with common nomenclature.

2. Enhance the clarity of Figure 7's caption by structuring it as (a), (b), and (c) preceding the figure description. For instance, start with "The paradigms of (a) cloud computing,"...

3. Modify the terminology in Table 4, replacing "AI Power" with a more suitable term such as "AI Processing Speed" for improved precision and clarity.

A final read-through for minor grammar errors is recommended for further refinement.

Author Response

Thank you for the feedback and suggestions. To address your concerns, we have made several important adjustments in response to your comments.

Comments and Suggestions for Authors:

This paper systematically reviews UAV-based real-time object detection, addressing the limited focus on onboard implementation in prior research. It offers a comprehensive hardware and software framework, enhancing its value as a reference for UAV real-time data processing. The paper also highlights the necessity of real-time detection in various application scenarios and investigates the impact of sensors, computing platforms, algorithms, and paradigms on accuracy and speed in UAV real-time object detection.

Overall, the paper is well-written; however, I recommend the authors conduct a final grammar check. Additionally, please consider the following suggestions for improvement:

Correct the terminology in Figure 5 to "Sankey diagram" rather than "Traffic diagram" to align with common nomenclature.

• I have used the Sankey diagram to replace the original description in Figure 5.
• The revision is highlighted in yellow, see in the caption of Figure 5.

Enhance the clarity of Figure 7's caption by structuring it as (a), (b), and (c) preceding the figure description. For instance, start with "The paradigms of (a) cloud computing,"...

• Thank you for this good advice, I have changed the figure description.
• The revision is highlighted in yellow, see in caption of Figure 7.

Modify the terminology in Table 4, replacing "AI Power" with a more suitable term such as "AI Processing Speed" for improved precision and clarity.

• I replaced AI power with AI performance, from the Nvidia official website.
• The revision is highlighted in yellow, see in Table 4.

Comments on the Quality of English Language:

A final read-through for minor grammar errors is recommended for further refinement.

• We have checked the grammar errors and made revisions.

Author Response File: Author Response.docx

Reviewer 2 Report

 

 

This paper aims at  systematically review real-time object detection from UAVS in terms of hardware selection, real- time detection paradigms, detection algorithms and their optimization technologies, and evaluation metrics. It is well-written and easy to follow. However, there are some concerns as follows:

1. The literature review needs improvement .

1.1 There are some relevant review papers missing such as :

1. Ramachandran, Anitha and Arun Kumar Sangaiah. “A Review on Object Detection in Unmanned Aerial Vehicle Surveillance.” International Journal of Cognitive Computing in Engineering (2021): n. pag.

The authors should add this and other missing ones and compare them and and what makes this study better?

1.2. There are some relevant papers missing such as :

1. Martinez-Alpiste, Ignacio et al. “Real-Time Low-Pixel Infrared Human Detection From Unmanned Aerial Vehicles.” Proceedings of the 10th ACM Symposium on Design and Analysis of Intelligent Vehicular Networks and Applications (2020): n. pag.(Realtime human detection with thermal imagery)

 

2. Golcarenarenji, Gelayol et al. “Illumination-aware image fusion for around-the-clock human detection in adverse environments from Unmanned Aerial Vehicle.” Expert Syst. Appl. 204 (2022): 117413.( Realtime human detection fusion of RGB and Thermal images)

 

Please add these and any other relevant missing papers to this review.

 

 

 

 

 

 

 

1. Minor editing of English language required. I would advise the authors to proofread the whole manuscript.

 

2. Some of abbreviations are not being defined throughout the manuscript (e.g. GPU, IMU,...). They should be all defined.

 

Author Response

Thank you for the feedback and suggestions. To address your concerns, we have made several important adjustments in response to your comments.

Comments and Suggestions for Authors:

This paper aims at systematically review real-time object detection from UAVS in terms of hardware selection, real- time detection paradigms, detection algorithms and their optimization technologies, and evaluation metrics. It is well-written and easy to follow. However, there are some concerns as follows:

1. The literature review needs improvement .

1.1 There are some relevant review papers missing such as :

1. Ramachandran, Anitha and Arun Kumar Sangaiah. “A Review on Object Detection in Unmanned Aerial Vehicle Surveillance.”International Journal of Cognitive Computing in Engineering(2021): n. pag.

The authors should add this and other missing ones and compare them and and what makes this study better?

• Thanks for sharing this paper, its now cited it in the manuscript. Compared to this article, in our article, a more detailed summary of how to implement real-time UAV detection is analyzed and discussed in terms of algorithms, hardware, and computing paradigms. It can provide reference and guidance for implementing real-time detection of UAVs.
• The revision is highlighted in yellow, see in line 78 -81.

1.2. There are some relevant papers missing such as :

1. Martinez-Alpiste, Ignacio et al. “Real-Time Low-Pixel Infrared Human Detection From Unmanned Aerial Vehicles.”Proceedings of the 10th ACM Symposium on Design and Analysis of Intelligent Vehicular Networks and Applications(2020): n. pag.(Realtime human detection with thermal imagery)
2. Golcarenarenji, Gelayol et al. “Illumination-aware image fusion for around-the-clock human detection in adverse environments from Unmanned Aerial Vehicle.”Expert Syst. Appl.204 (2022): 117413. ( Realtime human detection fusion of RGB and Thermal images)

Please add these and any other relevant missing papers to this review.

• The search terms were formulated at the beginning of the systematic review (Section 2.2). Based on these terms, we conducted a literature search in Web of Science and Scopus. The two articles you provided are indeed relevant to what is discussed in this paper but were not found through the terms and databases we set up. They could not be added to the review database due to the reproducibility of systematic reviews and to eliminate bias.

Comments on the Quality of English Language:

1. Minor editing of English language required. I would advise the authors to proofread the whole manuscript.

• We have checked the grammar errors and made revisions.

2. Some of abbreviations are not being defined throughout the manuscript (e.g. GPU, IMU,...). They should be all defined.

• Thanks for your suggestion, we have checked the abbreviations in the manuscript, and added the definitions.
• The revision is highlighted in yellow. See in caption of Table 4, and line 619 – 620.

 

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

The comments have been addressed. No further comments.

Minor editing of English language required.