Overview of Image Datasets for Deep Learning Applications in Diagnostics of Power Infrastructure

Round 1

Reviewer 1 Report (New Reviewer)

The study is an interesting one. I have only a few suggestions.

1. Major contribution of the study may be highlighted.

2. Research gaps may be elaborated more.

3. It will be better if the authors can incorporate prisma diagram.

4. Future research directions may be added.

5. Present research challenges may be described more.

Author Response

1. Major contribution of the study may be highlighted.

Thank you for your insightful comments and for raising this point. We followed this advice and introduced the following fragment to the manuscript introduction section:

The main idea behind this paper is to collect the datasets related to power lines, which are dedicated, among others, to improve the maintenance processes. The article tries to fill the gap in the field of datasets usable for training the deep learning models for the power industry. The overview provided here could be used as a source list for other researchers interested in training their own deep learning models with open source datasets.

 

2. Research gaps may be elaborated more.

We followed this suggestion and introduced this fragment into the manuscript:

The power lines are a specific object that generates additional complications during data collection, which are the most problematic: safety concerns (working with power lines and related infrastructure can be dangerous. It involves working at heights and in close proximity to high-voltage electrical systems, posing risks to workers and vehicles collecting data), private property and permits (power lines often cross private property, and it can be difficult to obtain permission from landowners or utility companies to access their land for data collection), variability of infrastructure (power line infrastructure can vary widely in design, age, and materials used. Creating a dataset that covers this variability is essential for it to be comprehensive and applicable to different scenarios), environmental factors (data collection for power line datasets can be affected by various environmental factors, such as weather conditions, e.g., rain, snow, or extreme temperatures), and natural events (e.g., storms or earthquakes, which can complicate the process).

 

3. It will be better if the authors can incorporate prisma diagram.

Thanks for this advice. We developed and appended such a diagram to the manuscript (Fig.1.).

 

4. Future research directions may be added.

We agree to that suggestion. We enumerated such a potential following actions in the “Conclusion” section (starting from paragraph 290).

 

5. Present research challenges may be described more.

Thank you also for this comment. We appreciate it and included a few new paragraphs to the introduction section, to address this issue. The following lines have been appended:

Currently, the field of power line dataset generation is evolving and researchers are facing several challenges to improve data quality, accessibility and utility. In our view, the most important current research challenges are:

- Data fusion and integration: Researching methods to combine data from multiple sensors and sources to create more informative datasets. Integrating data from different types of sensors, such as visual cameras, thermal imaging, and acoustic sensors, can provide a more comprehensive understanding of power line conditions,

- Real-time analysis and decision making: Advancing real-time data analytics to enable proactive decision making and predictive maintenance for power line assets. This includes developing algorithms that can quickly process large amounts of data to identify potential problems and optimize grid performance,

- Automated data collection: Develop advanced techniques for automated data collection from power line assets using drones, satellites, LiDAR, or other remote sensing technologies. Automation can reduce human intervention, improve data coverage, and enable more frequent data updates.

 

Once again, we would like to thank you for the constructive comments and suggestions, they helped us improve the paper substantially.

Reviewer 2 Report (New Reviewer)

Review of the Manuscript sensors – 2504957-  „The overview of image datasets for deep learning applications in diagnostics of power infrastructure” for the Sensors.

 

General Comments

            From my point of view, it is a very interesting topic and simultaneously it seems that to the best of my knowledge is the first empirical research provides two main parts. The first one presents information about data sets used in machine learning, especially deep learning. The other one of the review also discusses the use of the original data set containing 2630 high-resolution labeled images of power line insulators and comments on the potential applications of this collection.

The paper consists of the following sections: Introduction, Review and Conclusions.

However, I find some recommendations:

1.       It would be very useful to add in the "Introduction" section the purpose, objectives and hypothesis of the research.

2.       Also,  we consider the literature is not enough and that is why, we recommend the authors to refer to other recent works indexed in Web of Science, Scopus, Emerald, Cambrige, and of course MDPI Journals. We suggest that the authors cite papers published in MDPI journals and Web of Science Journals, such as:

 

1.       Vancea D.PC., Aivaz K.-A., Simion L., Vanghele D., (2021), Export expansion policies. An analysis of Romanian exports between 2005-2020 using Principal Component Analysis method and short recommandations for increasing this activity, Transformation in Business & Economics, No 2A (53A), pp.614-634, ISSN 1648-4460

 

2.       Aivaz K.-A., Munteanu I. F., Stan M.I., Alina Chiriac, (2022), A Multivariate Analysis on the Links Between Transport Noncompliance and Financial Uncertainty in Times of COVID-19 Pandemics and War, Sustainability, vol. 14, Issue 16, eISSN: 2071-1050, DOI:10.3390/su141610040

3.       Batrancea, L.M.; Tulai, H., (2022), Thriving or Surviving in the Energy Industry: Lessons on Energy Production from the European Economies. Energies, 15(22), 8532. https://doi.org/10.3390/en15228532.

 

The conclusions must be expanded with possible economic policy implications of the research undertaken.

All in all, I consider that the paper must be improved. As a result, the article can be published in the prestigious Sensors journal after major revisions.

Author Response

General Comments

            From my point of view, it is a very interesting topic and simultaneously it seems that to the best of my knowledge is the first empirical research provides two main parts. The first one presents information about data sets used in machine learning, especially deep learning. The other one of the review also discusses the use of the original data set containing 2630 high-resolution labeled images of power line insulators and comments on the potential applications of this collection.

The paper consists of the following sections: Introduction, Review and Conclusions.

However, I find some recommendations:

1. It would be very useful to add in the "Introduction" section the purpose, objectives and hypothesis of the research.

Thank you for your insightful comments and for raising this point. We appreciate it and included a few new paragraphs to the introduction section, to address this issue. The following lines have been appended:

As the main goal of the article is to present the process of creating datasets used in training deep learning models in the context of atypical objects such as high-voltage power lines. The process of creating datasets for extensive objects that are critical elements of the country's economy encounters several issues, as indicated in the presented paper. To achieve this, the following several conditions are recommended:

• Comprehensiveness assessment: To evaluate the comprehensiveness of existing datasets available on the field of the power line maintenance. This objective aims to identify any gaps or limitations in the current datasets and highlight areas where additional data collection may be needed.
• Data comparison: To compare multiple datasets from different sources or studies, aiming to identify similarities, differences, and potential inconsistencies. This objective can help researchers understand the strengths and weaknesses of different datasets and choose the most appropriate one for their analysis.
• Data accessibility and open data evaluation: To assess the availability and accessibility of datasets for the research community. This objective is relevant for promoting open data practices and making datasets easily accessible to other researchers.
• High-quality evaluation: To assess the quality and reliability of datasets by examining their data sources, collection methods, and data processing techniques. This objective helps ensure that the datasets used in research are of high quality and suitable for analysis.

 

2. Also,  we consider the literature is not enough and that is why, we recommend the authors to refer to other recent works indexed in Web of Science, Scopus, Emerald, Cambrige, and of course MDPI Journals. We suggest that the authors cite papers published in MDPI journals and Web of Science Journals, such as:

We took this opportunity and included several additional, currently issued sources and referenced to them.

 

The conclusions must be expanded with possible economic policy implications of the research undertaken.

We followed this advice and introduced the following fragment to the manuscript conclusions section:

In the long term, the economic implications of the study conducted will allow the development of tools to optimize the planning and maintenance of critical infrastructure through the early identification of areas of increased degradation, which will directly lead to an increase in the life of the infrastructure. Undoubtedly, in the short term, the implementation of advanced diagnostic technologies may require significant investments, which is a challenge for operators and energy companies.

 

All in all, I consider that the paper must be improved. As a result, the article can be published in the prestigious Sensors journal after major revisions.

 

Once again, we would like to thank you for the constructive comments and suggestions, they helped us improve the paper substantially.

Reviewer 3 Report (New Reviewer)

The contribution of this paper not clear and motivation 

where the application of Deep learning and where the techniques of DL

The title of the paper need to revised 

why Figure 1 and 2 doesn't contain 2023

name of section 2 not clear and doesn't reflect the information in this section

what is the information given from this sentence "A number of the listed image batches could be used for classification problems [9,11,13, 87 15–17,19,20,24–26,28–31]"

Table 1 is very long 3 pages

 

 

 

Author Response

The contribution of this paper not clear and motivation 

where the application of Deep learning and where the techniques of DL

Thank you for your insightful comments and for raising this point. We appreciate it and included a few new paragraphs to the introduction section, to address this issue. The following lines have been appended:

The main idea behind this paper is to collect the datasets related to power lines, which are dedicated, among others, to improve the maintenance processes. The article tries to fill the gap in the field of datasets usable for training the deep learning models for the power industry. The overview provided here could be used as a source list for other researchers interested in training their own deep learning models with open source datasets. […]

As the main goal of the article is to present the process of creating datasets used in training deep learning models in the context of atypical objects such as high-voltage power lines. The process of creating datasets for extensive objects that are critical elements of the country's economy encounters several issues, as indicated in the presented paper. To achieve this, the following several conditions are recommended:

• Comprehensiveness assessment: To evaluate the comprehensiveness of existing datasets available on the field of the power line maintenance. This objective aims to identify any gaps or limitations in the current datasets and highlight areas where additional data collection may be needed.
• Data comparison: To compare multiple datasets from different sources or studies, aiming to identify similarities, differences, and potential inconsistencies. This objective can help researchers understand the strengths and weaknesses of different datasets and choose the most appropriate one for their analysis.
• Data accessibility and open data evaluation: To assess the availability and accessibility of datasets for the research community. This objective is relevant for promoting open data practices and making datasets easily accessible to other researchers.
• High-quality evaluation: To assess the quality and reliability of datasets by examining their data sources, collection methods, and data processing techniques. This objective helps ensure that the datasets used in research are of high quality and suitable for analysis.

 

The title of the paper need to revised 

Thank you for prompting us to reconsider the title. Since we decided to focus mainly on decent datasets for the power industry domain, we also decided to stay with the current version of the title that servers its main purpose.

 

why Figure 1 and 2 doesn't contain 2023

We agree that it is important to provide the most current figures. Since the data for the current year is not complete, we have not included it in the chart to avoid misinterpretation.

 

name of section 2 not clear and doesn't reflect the information in this section

Thank you for this suggestion. We updated the title of the mentioned section.

 

what is the information given from this sentence "A number of the listed image batches could be used for classification problems [9,11,13, 87 15–17,19,20,24–26,28–31]"

This sentence collects all the record references that could be used to classify the power line elements. The formatting is induced by the TEX template we are using for this paper, which was provided by the journal.

 

Table 1 is very long 3 pages

Thank you for this advice. The mentioned table is the main asset of this paper, as it provides our readers with details about the studied datasets. We also do not want to force the readers to look for this information in several places, so we have compiled this rather long table. Since we have some conflicting options to follow, to split this table or to leave it as it is, we decided to take the latter option.

 

We also added several enhancements to the manuscripts suggested by other reviewers, including additional references and tables. Once again, we would like to thank you for the constructive comments and suggestions, they helped us improve the paper substantially. 

Reviewer 4 Report (New Reviewer)

This article is not suitable for publication. Some significant changes need to be made:

1. Authors should define: purpose, objectives, limitations, research questions of the article. It is not clear from the introduction exactly what the main task of this research is.

2. I recommend that the Review section be renamed to Literature Review or Related Work. In this section, the research questions should be answered.

3. Since Tables 1 and 2 are too long, I recommend that they be separated as Appendix 1 and cited in the text. The main text is now difficult to read.

4. It would be good to add a new Methodology, Experiment or similar section to demonstrate the practical application of the described datasets.

5. The authors mention machine learning algorithms in the introduction, but it is not clear how they are implemented and why they are described. It is good to experiment by applying one or more of the algorithms described in the introduction to the studied data sets. Then draw conclusions about the applicability of both algorithms and data sets.

Author Response

This article is not suitable for publication. Some significant changes need to be made:

1. Authors should define: purpose, objectives, limitations, research questions of the article. It is not clear from the introduction exactly what the main task of this research is.

Thank you for your insightful comments and for raising this point. We appreciate it and included a few new paragraphs to the introduction section, to address this issue. The following lines have been appended:

The main idea behind this paper is to collect the datasets related to power lines, which are dedicated, among others, to improve the maintenance processes. The article tries to fill the gap in the field of datasets usable for training the deep learning models for the power industry. The overview provided here could be used as a source list for other researchers interested in training their own deep learning models with open source datasets. […]

As the main goal of the article is to present the process of creating datasets used in training deep learning models in the context of atypical objects such as high-voltage power lines. The process of creating datasets for extensive objects that are critical elements of the country's economy encounters several issues, as indicated in the presented paper. To achieve this, the following several conditions are recommended:

• Comprehensiveness assessment: To evaluate the comprehensiveness of existing datasets available on the field of the power line maintenance. This objective aims to identify any gaps or limitations in the current datasets and highlight areas where additional data collection may be needed.
• Data comparison: To compare multiple datasets from different sources or studies, aiming to identify similarities, differences, and potential inconsistencies. This objective can help researchers understand the strengths and weaknesses of different datasets and choose the most appropriate one for their analysis.
• Data accessibility and open data evaluation: To assess the availability and accessibility of datasets for the research community. This objective is relevant for promoting open data practices and making datasets easily accessible to other researchers.
• High-quality evaluation: To assess the quality and reliability of datasets by examining their data sources, collection methods, and data processing techniques. This objective helps ensure that the datasets used in research are of high quality and suitable for analysis.

 

2. I recommend that the Review section be renamed to Literature Review or Related Work. In this section, the research questions should be answered.

Thank you for this suggestion. We updated the title of the mentioned section.

 

3. Since Tables 1 and 2 are too long, I recommend that they be separated as Appendix 1 and cited in the text. The main text is now difficult to read.

Thank you for this advice. The mentioned table is the main asset of this paper, as it provides our readers with details about the studied datasets. We also do not want to force the readers to look for this information in several places, so we have compiled this rather long table. Since we have some conflicting options to follow, to split this table or to leave it as it is, we decided to take the latter option.

 

4. It would be good to add a new Methodology, Experiment or similar section to demonstrate the practical application of the described datasets.
5. The authors mention machine learning algorithms in the introduction, but it is not clear how they are implemented and why they are described. It is good to experiment by applying one or more of the algorithms described in the introduction to the studied data sets. Then draw conclusions about the applicability of both algorithms and data sets.

We studied both comments and address them together. The main goal of the article is not focused on deep learning methods but rather on aggregating information and attempting to compare available datasets, saving time for other researchers. The collected material can be used for the detection and classification of elements of the energy infrastructure, mainly insulators.

 

We also added several enhancements to the manuscripts suggested by other reviewers, including additional references and tables. Once again, we would like to thank you for the constructive comments and suggestions, they helped us improve the paper substantially.

Round 2

Reviewer 2 Report (New Reviewer)

I agree this article's form.

Reviewer 4 Report (New Reviewer)

The authors made all changes that was recommended in the previous review.

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.

Round 1

Reviewer 1 Report

The authors need point out what problems are easily encountered or common in the actual testing process at present, and how much their research value is.

Author Response

General comments
The authors need point out what problems are easily encountered or common in the actual testing process at present, and how much their research value is.

Response:
We would like to thank the Editor and Reviewer for considering the paper and for the very insightful suggestions. We have refined the paper carefully following the posted remark. Please find below our response to this comment.

We agree that the motivation for this part was indeed too short. In the revised version of the manuscript, as suggested, we have added the following fragment to the 2.1. section to give the reader more details about the proposed aspect of the 2.1. section:

The essence of the study is detailed in two main artifacts: Table 1, which describes available data collections examined for this paper, and Table 2, which presents selected studies benefiting from these data. It should be noted that the primary focus of the collections is on the main elements of overhead power lines, such as conductors, towers, and insulators, since they are also the most frequently required interventions. From the point of view of maintenance management and the need for preventive measures to avoid failures in the transmission network, it is essential to develop appropriate data sets to support further research and development of algorithms, i.e. for detection or classification of:

• auxiliary equipment of overhead power lines:
– clamps for attaching wires and auxiliary lines,
– accessories intended for connecting clamps with insulators,
– accessories for suspending insulator chains on poles and connecting multi-row chains,
– protective equipment fixed at both ends of the insulator chains, designed to spread the electric field strength,
– power line spacers used to fix bundle wires and ensure a constant distance between them,
– tuned mass dampers (Stockbridge dumpers),
– protective marking (aviation obstruction signs, bird flight diverters),
– de-icing and anti-icing equipment,

• objects colliding with high voltage lines detection:
– branches and trees (problem of transmission corridors),
– stork nests on poles.

 

We also introduced another Figure depicting the aggregated view of the review and some additional comments on the conclusions.

 

Once again, we would like to thank you for your constructive comments and suggestions - they have helped us improve the paper significantly.

Best Regards

Bogdan Ruszczak
Pawel Michalski
Michal Tomaszewski

Reviewer 2 Report

This manuscript presents information on data sets that are used in the power line inspection process by applying machine learning techniques. Selected repositories of various data collections related to electrical infrastructure objects were compared.

--- The approaches, techniques and methodologies of the works that are part of the survey carried out are not sufficiently explained to justify a contribution to the state of the art.

--- The impact of the work presented here is not clearly appreciated or described, in the conclusions it only dedicates a couple of lines to its importance: "It can be used as part of the preparation … as a preliminary classifier in the sample annotation process."

--- In the appreciation of the present reviewer, the data collection described in section 2.1, which is the work of the same authors, disagrees with the line followed by this manuscript, since the methodologies of the other works referenced in the survey are not described with the same level of detail.

--- Since the observations indicated above imply the conformation of a work that would become largely different from the one presented here, this reviewer does not accept the manuscript for publication.

Author Response

General comments
This manuscript presents information on data sets that are used in the power line inspection process by applying machine learning techniques. Selected repositories of various data collections related to electrical infrastructure objects were compared.

Response:

Thank you for appreciating our work and for providing us with insightful comments. We have addressed all of them-we gather our point-to-point responses below.

 

Detailed comments and responses

1. The approaches, techniques and methodologies of the works that are part of the survey carried out are not sufficiently explained to justify a contribution to the state of the art.

Response:
Thank you for this comment. We have rewritten the main section of the paper - the subsection 2.1, and supplemented it with comments regarding this suggestion. Subsection 2.1 has been extended with two additional figures (Fig.2 and 3), and subsection 2.2. with the following text:

Developing a domain data collection, which was the main goal for this review, is rather difficult, especially when it should target a power industry equipment. Such a collection should allow one to deal with a narrow spectrum of problems and aid dedicated algorithms and applications development and encapsulate a narrow field knowledge. Since the most interesting elements of the power line infrastructure are objects located in its vicinity, it often requires data acquisition process realized in an upper section of the power poles. The training data recorded from the ground would be very different from that recorded during diagnostic flights. This creates an additional difficulty in the process of collecting good quality samples, as it usually forces the need to fly a helicopter in the vicinity of the power line, while still popular manual diagnostic processes. Such an approach significantly increases the cost of data set development and often excludes smaller organizations and individuals from the market. Another barrier to the creation of such domain data sets is the need for the expertise required to classify and assess damage to elements of towers or other power line equipment after data collection. As part of the review, the current state of available collections, their quality and usefulness from the perspective of power line diagnostics automation has been presented. In our opinion, the most valuable among the presented collections is the set [17], although it is not characterized by the highest number of training samples, it has all the elements that affect the quality and usefulness of this type of sets, and these are:

• Size: the collection should provide enough data to allow accurate power
line diagnostics. The more data, the better the chance of detecting damage. However, for the mentioned set [17], its size is not at the top of the list, suggesting that following the size criteria alone could be ambiguous. Its size is relatively small compared to others, but it is currently the most numerous in terms of found and labeled damage.
• Diversity: the collection should include a variety of data, different power line conditions, and address realistic scenes. In this way, it should be possible to identify problems in a variety of conditions, which is crucial for effective power line monitoring. In the set we have emphasized different groups of insulators and their damages have been collected (including two classes of damages: broken insulator shell, and flashover damaged insulator shell).

• Data quality: the image material in the selected set should provide the highest possible resolution to record, which later will help the precise marking process. This is one of the most important aspects of a well-prepared collection, while incorrect or inaccurate annotations can be misleading.

• Genericity: The method of registration partially disqualifies the usefulness of some sets for applications in diagnostic solutions based on air flights. In the promoted set, the acquisition was performed from the air, which refers to similar recordings during a scheduled diagnostic flight.

2. The impact of the work presented here is not clearly appreciated or described, in the conclusions it only dedicates a couple of lines to its importance: ”It can be used as part of the preparation . . . as a preliminary classifier in the sample annotation process.”

Response:
We agree that the motivation for this part was indeed too short. In the revised version of the manuscript, as suggested, we have added the following fragment to the 2.1. section to give the reader more details about the proposed aspect of the 2.1. section:

The essence of the study is detailed in two main artifacts: Table 1, which describes available data collections examined for this paper, and Table 2, which presents selected studies benefiting from these data. It should be noted that the primary focus of the collections is on the main elements of overhead power lines, such as conductors, towers, and insulators, since they are also the most frequently required interventions. From the point of view of maintenance management and the need for preventive measures to avoid failures in the transmission network, it is essential to develop appropriate data sets to support further research and development of algorithms, i.e.
for detection or classification of:
• auxiliary equipment of overhead power lines:
– clamps for attaching wires and auxiliary lines,
– accessories intended for connecting clamps with insulators,
– accessories for suspending insulator chains on poles and connecting multi-row chains,
– protective equipment fixed at both ends of the insulator chains, designed to spread the electric field strength,
– power line spacers used to fix bundle wires and ensure a constant distance between
them,
– tuned mass dampers (Stockbridge dumpers),
– protective marking (aviation obstruction signs, bird flight diverters),
– de-icing and anti-icing equipment,
• objects colliding with high voltage lines detection:
– branches and trees (problem of transmission corridors),
– stork nests on poles.

 

3. In the appreciation of the present reviewer, the data collection described in section 2.1, which is the work of the same authors, disagrees with the line followed by this manuscript, since the methodologies of the other works referenced in the survey are not described with the same level of detail.

Response:
We appreciate your suggestion for this section. The mentioned text was indeed too brief in the original version of the manuscript, and did not sufficiently bind these paper parts. We have elaborated on this in Section 2 and its sub-sections. We have also restructured this section, dividing the text into two main parts (subsections 2.1 and 2.2) and adding an introductory paragraph.
We have also expanded the description of the set in Table 1 (which was detailed in Subsection 2.2) and provided an additional figure there.

4. Since the observations indicated above imply the conformation of a work that would become largely different from the one presented here, this reviewer does not accept the manuscript for publication.

Response:
Since this part is crucial for the entire review, we have added additional paragraphs to this section, which was also suggested by other reviewers.
This section has also been expanded to include Figure 2, which provides additional aggregated information about the records reviewed.
We also asked our native-speaking colleague to proofread the entire manuscript. We are sure that thanks to your efforts, the manuscript is much more readable and clearer, and the main message we wanted to convey has been corrected.

 

 

Once again, we would like to thank you for your constructive comments and suggestions - they have helped us improve the paper significantly.

Best Regards

Bogdan Ruszczak
Pawel Michalski
Michal Tomaszewski

Reviewer 3 Report

Review in the field is neccesary. So topic of the paper is interesting.

There is lack of connection between first part of the paper and subchapter 2.1. Authors should more explain the connection.

Structure of the paper should be improved. Currently there is introduction, review and conclusion, but inside chapter 2 is one subchapter 2.1. I suggest to divide chapter 2 into two subchapters. It will be better for structure of the paper.

 

Author Response

General comments
Review in the field is neccesary. So topic of the paper is interesting.

Response:
Thank you for appreciating our work and for providing us with insightful comments. We have addressed all of them - we gather our point-to-point responses below.

Detailed comments and responses
1. There is lack of connection between first part of the paper and subchapter 2.1. Authors should more explain the connection. The authors need point out what problems are easily encountered or common in the actual testing process at present, and how much their research value is.

Response:
Thank you for this comment - the mentioned description was indeed too brief in the original version of the manuscript and not sufficiently tied up those paper parts. We have elaborated on that in more detail in the section 2 and its subelements.

To increase this part reception we added another Figure there (Fig. 2)
We also appended the following to the manuscript to extend the comments in the section 2:

The following part of the paper consists of two subsections. The first provides an overview of the datasets with an indication of their main features and discusses a compilation of studies that use such data to develop algorithms and applications for the power industry. The second subsection includes a consideration of an example collection, which details the structure and potential applications of such collections, and deals with the detection of power insulators in images.

2. Structure of the paper should be improved. Currently there is introduction, review and conclusion, but inside chapter 2 is one subchapter 2.1. I suggest to divide chapter 2 into two subchapters. It will be better for structure of the paper.

Response:
We appreciate your suggestion. We have restructured the manuscript accordingly.
Section 2 has been split into two subsections and the additional introduction to this section has been added.
Since this part is crucial to the entire review, we have added additional paragraphs to this section, as suggested by other reviewers.

We also asked our native-speaking colleague to proofread the entire manuscript. We are sure that thanks to your efforts, the manuscript is much more readable and clearer, and the main message we wanted to convey has been corrected.

 

Once again, we would like to thank you for your constructive comments and suggestions - they have helped us improve the paper significantly.

 

Best Regards

Bogdan Ruszczak
Pawel Michalski
Michal Tomaszewski

Round 2

Reviewer 1 Report

This paper does not provide a detailed and in-depth analysis of the specific field, only 50 articles, and most of the content of the article is in the column dataset. So my comment is to reject this paper.

Reviewer 2 Report

The authors of this manuscript present significant changes in the preparation of their results, tables, and applied methodology, for which reason this reviewer considers that the manuscript in its current form is suitable for publication.