Next Article in Journal
A Low-Complexity Security Scheme for Drone Communication Based on PUF and LDPC
Previous Article in Journal
Achieving Robust Learning Outcomes in Autonomous Driving with DynamicNoise Integration in Deep Reinforcement Learning
 
 
Article
Peer-Review Record

Generating 3D Models for UAV-Based Detection of Riparian PET Plastic Bottle Waste: Integrating Local Social Media and InstantMesh

by Shijun Pan 1,*, Keisuke Yoshida 1, Daichi Shimoe 1, Takashi Kojima 2 and Satoshi Nishiyama 1
Reviewer 1: Anonymous
Reviewer 2:
Reviewer 3: Anonymous
Submission received: 2 August 2024 / Revised: 30 August 2024 / Accepted: 5 September 2024 / Published: 9 September 2024

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

In” Generating 3D Models for UAV-based Detection of Riparian PET Plastic Bottle Waste: Integrating Local Social Media and InstantMesh” the authors developed a data generation approach by combining social media and single-image-based 3D model generation algorithms . The proposed method uses the open source model InstantMesh as the core to generate a dataset that can be trained. This work also uses YOLOv8 to verify the usability of the dataset. This work provides a good data generation idea for small target recognition, especially for recognition in remote sensing images.

1. This paper solves the problem of data mismatch between waste management and local environment, and develops a method to generate 3-D datasets for local environment.

2. In the process of dataset generation, this work takes into account the size of the target in the drone image, which is reasonable and makes the generated data more effective.

Given the points above, I think this work is innovative enough, but it still needs to be improved in the following aspects

1. This work is centered on InstantMesh, but the description of InstantMesh in this article is limited and insufficient to explain the reason for using InstantMesh.

2. For the generated data set, this article uses YOLOv8 to verify the feasibility, but it is necessary to use multiple target detection algorithms for verification. The author can try multiple algorithm verifications to prove that the algorithm for data generation is feasible.

3. The presentation of the training results in Fig16 and Fig24 lacks description. I personally think that these results do not need to be presented, or only some of the results need to be presented, but the description of the results is the key point.

4. I still want to know if the author can design experiments to prove the versatility of the data generation method in this work, such as whether the data generated by this method cannot be used or the effect is not good when we change the location.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The manuscript by Pan et al. presents a very interesting method for macro-litter detection from photos, which could have many practical applications. I suggest publication after major revision. In my review, I have recommended a series of corrections, mostly related to better structuring and clarifying the paper.

General Comments:

The paper lacks answers to the following two questions:

  1. Can the method be used only for training algorithms targeted at monitoring macroplastic in flowing water, or can it also be applied to plastics trapped on riverbanks (e.g., on or inside vegetation)?
  2. What is the impact of seasonal changes in vegetation on the method’s accuracy and general utility in riverbank environments?

Specific Comments on Sections and Figures:

Abstract:

  • The aims of the study should be more precisely formulated in the first part of the abstract.
  • Please add one sentence describing specific results obtained by your study.
  • Please include a sentence describing the direct practical implications of your work. A brief version of this sentence, similar to what is found in lines 117-120, would fit perfectly here.

Introduction:

  • I suggest concluding the Introduction section at line 120.
  • The paragraphs from lines 121-205 should be moved to a separate section titled "Materials and Methods."
  • Line 25-26: It is not clear which type of waste you are referring to. Please specify and add key references.
  • Line 26-27: Please add a reference to support this statement.
  • Line 29: I think that some work describing the risk of PET bottle fragmentation and resulting secondary microplastic formation in rivers should be cited at the end of this sentence. 
  • Lines 29-32: I suggest adding references containing information that plastic can also be fragmented before it reaches the ocean, e.g., in rivers.
  • Line 33: Please replace "heavy toll" with a more precise term.
  • Figure 1: I generally like the layout of this figure. However, adding some general descriptions of the environments (e.g., rivers, oceans) and processes (e.g., transport, fragmentation) presented in the circles will help readers understand what you mean.
  • Line 40: More general references should be cited here.
  • Line 70-71: Please specify whether this company detects and categorizes individual litter objects or dumping site locations.
  • Line 75: This statement is a bit vague. Please revise or remove it.
  • Line 77-79: Please correct the style of this sentence.
  • Line 83-86: Do these patrols use drones, or only visual inspection?

Figures 12 and 13: I suggest moving these figures to the supplementary materials. Tables 1-5: I suggest moving these tables to the supplementary materials.

Results:

  • The results section is written in very technical language. Please add introductory sentences before the technical statements to facilitate reader understanding.
  • Figure 16: Please add more detailed descriptions of the data presented in this figure.
  • Figures 17 and 18: As above.
  • Figure 18: Please increase the size and resolution of this figure.
  • Figure 19: Please remove the bolding from the axis numbers.
  • Lines 322-326: Please revise these sentences for clarity.
  • Figures 21 and 22: Please move these figures to the supplementary materials.
  • Figures 25, 26, 27: Please move these figures to the supplementary materials.

Discussion:

  • Before the technical discussion, please add a general sentence describing a key result achieved in this work.
  • The discussion is too brief. All parts of the results section currently presented as discussion should be moved here and extended.
  • Limitations and ideas for future improvements should be carefully presented in the revised discussion section.
  • Lines 375-376: I don’t understand this statement.

Conclusion:

  • The conclusion should be completely revised to be more general and clear for a broad audience. Please avoid technical terms and provide only the key results and main take-home message from your work.

 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

I consider that the work presented by the authors represents a relevant contribution and application to problems of a global nature. However, the following comments should be addressed:

1.-In the introduction they should highlight the main contribution or contributions of their work, since at first glance they cannot be visualized.

2.-When analyzing PET images, could you classify if the PET bottle is dirty or only the algorithm detects the existence of the bottle? This would be useful for recycling.

3.- The presented conclusion should be enriched.

4. Diagrams should be made to better understand the designed approach.

5. In the References section, several references are incomplete

Comments on the Quality of English Language

I consider that the work presented by the authors represents a relevant contribution and application to problems of a global nature. However, the following comments should be addressed:

1.-In the introduction they should highlight the main contribution or contributions of their work, since at first glance they cannot be visualized.

2.-When analyzing PET images, could you classify if the PET bottle is dirty or only the algorithm detects the existence of the bottle? This would be useful for recycling.

3.- The presented conclusion should be enriched.

4. Diagrams should be made to better understand the designed approach.

5. In the References section, several references are incomplete

 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

All my concers have been addressed. 

 

Reviewer 3 Report

Comments and Suggestions for Authors

Considering the final version of the article presented by the authors, they have addressed the comments made to the previous version resulting in a comprehensive work that is a valuable contribution to the state of the art. 

Comments on the Quality of English Language

Considering the final version of the article presented by the authors, they have addressed the comments made to the previous version resulting in a comprehensive work that is a valuable contribution to the state of the art. 

Back to TopTop