MineCam: Application of Combined Remote Sensing and Machine Learning for Segmentation and Change Detection of Mining Areas Enabling Multi-Purpose Monitoring
, Marcin Maksymowicz 3, Dariusz Tanajewski 3, Wojciech Kaczan 2,4, Maciej Zięba 1 and Marek Wilgucki 2Round 1
Reviewer 1 Report (Previous Reviewer 1)
Comments and Suggestions for AuthorsThis manuscript used machine learning techniques to segment images of mines of different types, raw materials and geographical locations. The massive availability of remote sensing data, multiple labels in the model setup can support the results of this study. The reviewer recommends that when defining labels and classes, it is important to clearly articulate the underlying rationale or provide citations from other paper. It is also recommended that the authors create a flowchart encompassing all the techniques and contents of the study. This will facilitate better understanding for the readers.
Some specific changes are as follows:
1. Line 41 “1.2. Challenges”
The authors summarize three challenges that need to be addressed. The reviewer feel that citing only one article to illustrate the shortcomings of the existing research is insufficient.
2. Lines124-125 “Based on the aforementioned considerations, we selected 10 classes of mining area coverage.”
Where did the authors select these 10 classes? The list provided above is not sufficiently detailed to encompass all potential identification labels that might be involved in the mining area.
3.Line133 “2.2. Subsets of training data”
The reviewer understands the title of section 2.2 to pertain to training data, yet the first paragraph elaborates on the use of models. It is recommended that these topics be presented separately for clarity.
4.Lines 144-145 “Consequently, we opted to train two multi-class segmentation models and two binary models.”
Lines 146-147 “As a result, three sets were created which grouped the labels into the following categories:”
What is the relationship between “two multi-class segmentation models and two binary models” and the four sets?
4.Line137 “e.g. TSFs or dumping grounds as in 3”
TSFs need to be written with both the full name and the abbreviation on its first appearance.
5.Line 483 “5. Discussion”
The reviewer suggests that in the Discussion section, the authors should compare the results of this study with those of existing studies for comprehensive analysis. Avoid using an “We believe that repeating similar studies on satellite data with better spatial resolution…(Line523)” expression, which is very uncritical.
Author Response
Please see the attachment
Author Response File: 
Author Response.pdf
Reviewer 2 Report (Previous Reviewer 2)
Comments and Suggestions for AuthorsThis is the revised version of manuscript submitted by the same authors earlier. In this revised version, the authors have made all changes suggested earlier. The authors argument on considering the data from 2016 to 2022 is acceptable based on the availability of satellite data.
All figures now appear in the revised manuscript. Further, the authors have clarified different models, namely model 1 to model 4 clearly in this version. The authors have argued in their rebuttal that different models are trained to improve overall performance. Further, the authors agree that the model may not be able to distinguish archeological or similar excavations that may have similar features. I suggest the authors to include these arguments in a separate section named limitations of the present study before conclusions section.
Overall, I recommend the manuscript in present form to be accepted for publication, subject to the above minor suggestions.
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Reviewer 3 Report (Previous Reviewer 4)
Comments and Suggestions for AuthorsI have no further comments. The authors improved the manuscript. The results are now clearly presented and they align with the derived conclusions.
Author Response
Please see the attachment
Author Response File: Author Response.pdf
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
Comments and Suggestions for AuthorsThe manuscript developed several machine learning models to identify mining areas, with all samples divided into three groups and corresponding to four models. However, all the figures in the manuscript are missing, making it difficult to assess the reliability and value of the research based on text alone. The authors need to provide the missing figures to support their findings.
Comments on the Quality of English LanguageThe quality of English writing is good, and the research content can be clearly expressed.
Author Response
please see the attachment
Author Response File: Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsPlease see the attached document for my review.
Comments for author File: Comments.pdf
Please check for typographical error in abstract section.
Author Response
please see the attachment
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsPoint 1: Kindly follow the style in writing the article title for MDPI journals.
Point 2: Kindly change ‘learninf’ to ‘learning’ in the title.
Point 3: Insert a comma between each author except for the two last authors.
Point 4: Use a different superscript if the authors use two departments and (or) affiliations.
Point 5: Do not use the first person (i.e., our, we, and the like) in the paper.
Point 6: What does ‘this’ refer to in lines 34, 47, 52, and other parts of the paper? Kindly be specific.
Point 7: Kindly follow the correct citation style (e.g., lines 36, 41, and in other parts of the paper).
Point 8: Kindly change ‘resulting’ to ‘resulted’ in line 47.
Point 9: Cite examples of common models and the land cover classes these models typically cover in line 50.
Point 10: Cite examples of these previous studies in line 58.
Point 11: Cite examples of these studies in line 63.
Point 12: Kindly define all acronyms and (or) abbreviations used during their appearance in the paper and use them consistently.
Point 13: All figures are missing.
Point 14: How was the differentiation stated in line 95 specifically performed?
Point 15: How do the authors define and (or) quantify small, medium, and large mining areas?
Point 16: Why only use one cloudless image yearly, as in line 99?
Point 17: The authors mentioned that reclaimed areas are one of the six common land cover classes for surface mining areas. Why did the authors exclude these areas from the study?
Point 18: Ensure that figures and tables appear immediately during their first mention in the paper.
Point 19: Besides those stated in lines 145–147, what other similarities do these labels have in common, making them one set?
Point 20: There are three sets stated in line 138. However, the authors mentioned that there is an additional fourth set. Kindly make sure the readers understand the total number of sets.
Point 21: There are four models stated in lines 160–161. Can the authors provide concise details about these models?
Point 22: How did the authors deal with the heterogeneity in line 162?
Point 23: How did the authors do the surface mine type classification stated in lines 187–189 when they needed to know the depths and (or) volumes of the excavations and the like?
Point 24: What are these non-overlapping subsets in line 207?
Point 25: How is the optimal band selected in concise detail?
Point 26: What are these data augmentation techniques in lines 229–230?
Point 27: Kindly provide more discussion about band configuration.
Point 28: What are the abbreviations in Table 3 and other tables?
Point 29: Follow the proper style in listing references.
Comments on the Quality of English LanguageModerate editing of the English language is required.
Author Response
please see the attachment
Author Response File: Author Response.pdf
Reviewer 4 Report
Comments and Suggestions for AuthorsThe authors propose a framework based on machine learning methods and remote sensing to derive suitable models for mining area segmentation.
The achieved results within the proposed scenarios are not outstanding BUT the authors deal with a very difficult problem for which many realistic variables intervene.
My personal opinion is that the work done is interesting and worthy of dissemination, and it may lay the ground for deeper studies to be conducted on the specific machine learning methods: how they can be modified/improved for example. The authors should address in forthcoming studies (if they plan any).
The main issue is that the manuscript presents some compilation error and therefore, none of the pictures was displayed. For this reason, the final evaluation of the manuscript cannot be fully addressed as the results that are mentioned to be in the figures cannot be seen.
Please, for the next submission, make sure the manuscript is correctly built before proceeding to the final stage of the submission.
Author Response
please see the attachment
Author Response File: Author Response.pdf
