Localization of Cracks in Concrete Structures Lacking Reference Objects and Feature Points Using an Unmanned Aerial Vehicle

Round 1

Reviewer 1 Report

This research proposed a new image-based concrete crack localization method. A video-based safety inspection experiment was conducted using unmanned aerial vehicles. However, the authors only provide the concept of crack localization method without any details of valuable processing procedure. The results are new but they cannot provide a valuable reference for the future research. Hence, it is suggested to revise the manuscript completely before it can be accepted.

(1) The writing of manuscript should be improved. The first sentence should be rewritten since it doesn’t make sense in logically. The abbreviations of UAV, SIFT should be shown in full title in the abstract, as well as in the body of the text. The authors had better not to decribe in the first person. The sentences in the abstract should not directly copy from the body of the text.

(2) The authors should provide the reference literatures for the two image processing techniques in lines 172-174 and for the three crack localization methods in lines 204-207 and interprete the ones used in the study with details.

(3) The concrete mathematic analysis process of image processing technique and crack localization method should be provided in the manuscript, not just with the process steps. It is far from being enough to interprete the concept, it seems to be not much good to the researchers. It's just more of an application.

It's good.

Author Response

Reviewer #1

→ We thank the editor and reviewers for their detailed comments and suggestions. We have thoroughly revised and improved our manuscript. Our responses to the reviewer comments are appended below.

 

1. The writing of manuscript should be improved. The first sentence should be rewritten since it doesn’t make sense in logically. The abbreviations of UAV, SIFT should be shown in full title in the abstract, as well as in the body of the text. The authors had better not to describe in the first person. The sentences in the abstract should not directly copy from the body of the text.

→ Thanks for the good comments. We have corrected and rewritten the first sentence. In addition, the abstract of the paper was thoroughly revised and rewritten. And the sentences described in the first person in the text have been corrected.

2. The authors should provide the reference literatures for the two image processing techniques in lines 172-174 and for the three crack localization methods in lines 204-207 and interprete the ones used in the study with details.

→ Thanks for the good comments. In sections 2.3 and 2.4, we describe the methods used in this study and provide references.

Section 2.3:

(a) Distorted image correction using homography matrix (Ref 33-34)

(b) Feature point extraction using SIFT algorithm (Ref 35)

Section 2.4:

(a) Construction of spatial information using image processing techniques (Refs 36-37)

(b) Merging of spatial information layers based on point cloud technique and image stitching technique (Ref 40 ~ 42)

(c) Crack localization and Data validation (methodology developed and proposed in this study)

In addition, the algorithms used in this study are open source algorithms commonly used in vision technology. This study is meaningful in developing application technology that can localization of concrete cracks by combining the corresponding element technologies.

 

3. The concrete mathematic analysis process of image processing technique and crack localization method should be provided in the manuscript, not just with the process steps. It is far from being enough to interprete the concept, it seems to be not much good to the researchers. It's just more of an application.

→ Thanks for the good comments. This study did not develop a new vision algorithm technology, but used a combination of previously developed element algorithms. Through this, the purpose of this study is to develop crack localization technology in the field of architecture and civil engineering. Therefore, in Chapter 2, we tried to write an explanation of the methodology for estimating the location of concrete cracks proposed in this study, and the contents of the paper were modified to provide a more detailed explanation.

Author Response File: Author Response.docx

Reviewer 2 Report

In the presented study, the authors reported on some efforts done for conduct a new image-based concrete crack location estimation study, seeking improvements in the existing methods. Image correction and feature point extraction were performed using the homography matrix and SIFT algorithm, and four cracks were defined using the point cloud technique, image stitching technique, and layer merging. The offered topic may be of interest to the journal readers, but the work should be revised before it can be considered further. The following are some of the concerns that must be addressed in the manuscript:

 

(1) In the abstract: The abstract part needs to be revised and modified to present only the most relevant goal, and idea, and get quickly to the main point of the paper. It is also essential to highlight work novelty. The authors should also check and reduce some unnecessary information in the abstract. There are some abbreviations in the abstract, I suggest the authors remove them.

(2) In the introduction, the authors tried to highlight the significance of the study, however, the novelty of the study should be emphasized further.

(3) In the “Materials and methods”, “2.1. Overview”, It is recommended to use a flowchart to express the content below “The proposed method comprises the following steps”.

(4) It is recommended to enlarge Figure 1 a. The current image quality is very low, and the corresponding content cannot be clearly expressed. The quality of other images should also be improved as much as possible.

(5) Figures 4 and 1a are duplicated. It is recommended to use a different expression by repeating two images in the same manuscript, which is not necessary. The following other images also have similar situations, and it is strongly recommended to make changes.

(6) The authors should leave a space between values and units (eg. RMSE X= 0.65 m instead of RMSE X= 0.65m). This applies to the other parts of the manuscript.

(7) The conclusion should be revised based on the material, the objective, the main results obtained, as well as clarification of the future applicability of the study.

Appropriate English modifications are necessary.

Author Response

Reviewer #2

→ We thank the editor and reviewers for their detailed comments and suggestions. We have thoroughly revised and improved our manuscript. Our responses to the reviewer comments are appended below.

 

1. In the abstract: The abstract part needs to be revised and modified to present only the most relevant goal, and idea, and get quickly to the main point of the paper. It is also essential to highlight work novelty. The authors should also check and reduce some unnecessary information in the abstract. There are some abbreviations in the abstract, I suggest the authors remove them.

→ We have thoroughly revised and rewritten the abstract of this paper. Thanks for the good comments.

 

2. In the introduction, the authors tried to highlight the significance of the study, however, the novelty of the study should be emphasized further.

→ We modified the contents by adding the novelty of the research and differentiation from the existing research in the introduction. Thanks for the good comments.

 

3. In the “Materials and methods”, “2.1. Overview”, It is recommended to use a flowchart to express the content below “The proposed method comprises the following steps”.

→ We provided Figure 1 as a visualized flow chart, and we have changed the title of the figure clearly. Thanks for the good comments.

 

4. It is recommended to enlarge Figure 1 a. The current image quality is very low, and the corresponding content cannot be clearly expressed. The quality of other images should also be improved as much as possible.

→ We have generally replaced the image quality of all figures in the paper (At least 300dpi). In addition, in the case of Figure 1, the schematic flow chart for this study was expressed as a visualization, and it was judged that it could be expressed sufficiently without enlarging it. Thanks for the good comments.

 

5. Figures 4 and 1a are duplicated. It is recommended to use a different expression by repeating two images in the same manuscript, which is not necessary. The following other images also have similar situations, and it is strongly recommended to make changes.

→ As mentioned in the reviewer's comment in answer 4, in the case of Figure 1, it is a picture to briefly explain the research flow chart in the form of visualization. Figures that are judged to be the same figure represent the research results of each chapter and have higher quality and resolution. Thanks for the good comments.

 

6. The authors should leave a space between values and units (eg. RMSE X= 0.65 m instead of RMSE X= 0.65m). This applies to the other parts of the manuscript.

→ All relevant contents in the paper have been corrected. Thanks for the good comments.

 

7. The conclusion should be revised based on the material, the objective, the main results obtained, as well as clarification of the future applicability of the study.

→ In the conclusion chapter, we added the applicability and research and development prospects of the methodology proposed in this study. Thanks for the good comments.

Author Response File: Author Response.docx

Reviewer 3 Report

SUMMARY

The article submitted for review is relevant. It considers the issue of localization of cracks in concrete structures lacking reference objects and feature points using an unmanned aerial vehicle. The authors raised an interesting problem. Since most crack investigation tasks are based on visual inspection, it has limitations in that it requires a lot of manpower and resources to carry out the inspection. The authors proposed a solution to the problem by replacing the visual inspection by personnel with image-based inspection using an unmanned aerial vehicle, image processing and deep learning technology. Thus, their study is practically significant, and at the same time, during the study, the authors received scientific novelty. The reviewer believes that the study was performed at a high level and presented an interesting topic. Thus, the authors' study can be used to improve traditional methods of visual inspection of infrastructure and facilities. All this will improve the quality of construction in general and preserve the existing infrastructure fund. The reviewer believes that the article should be supported, but at the same time, a number of comments should be corrected. The reviewer's comments are presented below.

 

COMMENTS

1. The authors in their abstract did not formulate the problem clearly enough. The well-known problem of replacing the workforce with more optimized methods, for example, unmanned aerial vehicles, is described in too much detail, but this fact is well known and does not require so much space in the abstract. At the same time, the authors did not identify a scientific problem, which may be the need to collect image data for interpretation, as well as scientific deficits that arise during the development of a new method. All this should be formulated as a scientific problem at the beginning of the abstract.

2. In general, the abstract should be slightly restructured to make it more consistent with the content of the article.

3. Not very well chosen keywords. In particular, the word "localization" or the word "crack" are not keywords for studies of this level. Authors should formulate phrases more specifically in relation to a particular article, and not some well-known single terms.

4. The introduction presented by the authors does not reflect the current state of the issue well enough. The reviewer would recommend considering various methods for detecting the location of cracks in concrete structures, perhaps using other methods, such as artificial intelligence, or using image analysis, or other methods that may be close to the method using unmanned aerial vehicles. Do not forget about traditional methods.

5. The reviewer believes that the 13 analyzed papers in the Introduction section do not quite reflect the scientific novelty. The number of analyzed references should be increased by 7-10 pieces, which will relate to the recognition of crack detection in concrete and concrete structures. In this case, preference should be given to works for the last 5 years. This will help make the article more relevant.

6. In the Materials and Methods section, I would like to see a flowchart of the study. The authors presented quite a lot of methodological material, but I would like to structure the study using a flowchart: what tasks and goal the authors set, what methods and materials were used for this, and what final result they went to. Such a flowchart will help the reader to better navigate the study.

7. The section "Experimental results" looks interesting. It contains many photos, but some of them are in low quality and poor resolution. Authors are encouraged to submit figures 6 and 7 in higher quality. The same remark applies to Figure 8. The results of the study are poorly discussed.

8. The "Discussion" section should be strengthened. Authors should compare their results in detail with those previously obtained by other authors. Here, a table looks appropriate that compares the results obtained using unmanned aerial vehicles with the results, for example, of other crack detection methods using a neural network, machine learning methods, other artificial intelligence methods, as well as traditional methods. This will help to understand the scientific contribution of the article to engineering and scientific practice.

9. Conclusions should be specified in terms of prospects for the development of the study, practical value and specific recommendations for application. At what facilities should such an unmanned aerial vehicle and such a method be used, in what areas of engineering facilities is this applicable?

10. It is also suggested to work a little with the References. The authors touch on a very modern method, but consider many works older than the last 5 years. Thus, the list of references should be supplemented with at least 15 references for the last 5 years, dedicated to the research topic.

11. In general, the article has some comments on the style of presentation and some inaccuracies in English.

12. The reviewer responds positively to this article and believes that after correcting the comments made, the article can be published in the journal Applied Sciences.

Comments for author File: Comments.pdf

Small language check needed.

Author Response

Reviewer #3

→ We thank the editor and reviewers for their detailed comments and suggestions. We have thoroughly revised and improved our manuscript. Our responses to the reviewer comments are appended below.

 

1. The authors in their abstract did not formulate the problem clearly enough. The well-known problem of replacing the workforce with more optimized methods, for example, unmanned aerial vehicles, is described in too much detail, but this fact is well known and does not require so much space in the abstract. At the same time, the authors did not identify a scientific problem, which may be the need to collect image data for interpretation, as well as scientific deficits that arise during the development of a new method. All this should be formulated as a scientific problem at the beginning of the abstract.

→ We have thoroughly revised the abstract of the paper based on the reviewer's feedback. Thanks for the good comments.

 

2. In general, the abstract should be slightly restructured to make it more consistent with the content of the article.

→ We have thoroughly revised and rewritten the abstract of this paper. Thanks for the good comments.

 

3. Not very well chosen keywords. In particular, the word "localization" or the word "crack" are not keywords for studies of this level. Authors should formulate phrases more specifically in relation to a particular article, and not some well-known single terms.

→ We have added more specific keywords to the existing ones, including image stitching and laser pointer, and modified it to concrete crack. Thanks for the good comments.

 

4. The introduction presented by the authors does not reflect the current state of the issue well enough. The reviewer would recommend considering various methods for detecting the location of cracks in concrete structures, perhaps using other methods, such as artificial intelligence, or using image analysis, or other methods that may be close to the method using unmanned aerial vehicles. Do not forget about traditional methods.

→ This study does not define cracks by detecting or identifying cracks. The necessity of this study has been reinforced as the majority of studies are focused on crack detection and identification, and research on crack localization is insufficient. In addition, the content has been modified so that readers can clearly understand that the purpose of this study is the development of crack localization technology. Thanks for the good comments.

 

5. The reviewer believes that the 13 analyzed papers in the Introduction section do not quite reflect the scientific novelty. The number of analyzed references should be increased by 7-10 pieces, which will relate to the recognition of crack detection in concrete and concrete structures. In this case, preference should be given to works for the last 5 years. This will help make the article more relevant.

→ We have added 10 references related to the detection, identification and location of concrete cracks within the last 5 years in the introduction chapter. Thanks for the good comments.

 

6. In the Materials and Methods section, I would like to see a flowchart of the study. The authors presented quite a lot of methodological material, but I would like to structure the study using a flowchart: what tasks and goal the authors set, what methods and materials were used for this, and what final result they went to. Such a flowchart will help the reader to better navigate the study.

→ We provided Figure 1 as a visualized flow chart, and we have changed the title of the figure clearly. Thanks for the good comments.

 

7. The section "Experimental results" looks interesting. It contains many photos, but some of them are in low quality and poor resolution. Authors are encouraged to submit figures 6 and 7 in higher quality. The same remark applies to Figure 8. The results of the study are poorly discussed.

→ We have generally replaced the image quality of all figures in the paper (At least 300dpi). Also, we supplemented and revised the discussion and conclusion chapters. Thanks for the good comments.

 

8. The "Discussion" section should be strengthened. Authors should compare their results in detail with those previously obtained by other authors. Here, a table looks appropriate that compares the results obtained using unmanned aerial vehicles with the results, for example, of other crack detection methods using a neural network, machine learning methods, other artificial intelligence methods, as well as traditional methods. This will help to understand the scientific contribution of the article to engineering and scientific practice.

→ By revising and supplementing the contents of the discussion chapter, we have presented a difference from previous papers related to the subject of this study. It also describes possible future applications. Thanks for the good comments.

 

9. Conclusions should be specified in terms of prospects for the development of the study, practical value and specific recommendations for application. At what facilities should such an unmanned aerial vehicle and such a method be used, in what areas of engineering facilities is this applicable?

→ In the conclusion chapter, we added the applicability and research and development prospects of the methodology proposed in this study. Thanks for the good comments.

 

10. It is also suggested to work a little with the References. The authors touch on a very modern method, but consider many works older than the last 5 years. Thus, the list of references should be supplemented with at least 15 references for the last 5 years, dedicated to the research topic.

→ As mentioned in the reviewer's comment in answer 5, we added 10 references in the introduction chapter. In addition, 5 references related to this research topic were added. Thanks for the good comments.

 

11. In general, the article has some comments on the style of presentation and some inaccuracies in English.

→ As we reviewed the contents of the paper again, we modified the parts that were explained in the first person. Also, the English of the text has been re-corrected throughout. Thanks for the good comments.

 

12. The reviewer responds positively to this article and believes that after correcting the comments made, the article can be published in the journal Applied Sciences.

→ I would like to express my gratitude to the reviewer for their positive response to our article. We are committed to addressing and correcting the comments provided by the reviewer to ensure that our article meets the standards for publication in the journal Applied Sciences. Your feedback is greatly appreciated, and we will work diligently to make the necessary revisions. Thanks for the good comments.

Author Response File: Author Response.docx

Round 2

Reviewer 3 Report

The authors did a good job of improving the manuscript and took into account all the comments of the reviewer. The reviewer has no more comments.

Author Response

Thank you for your positive feedback. We appreciate the thorough review and are pleased to hear that the improvements made to the manuscript have been well-received. If you have any further suggestions or questions in the future, please feel free to reach out. We look forward to the possibility of our manuscript being accepted for publication in Applied Sciences.