The Use of Machine Learning and Satellite Imagery to Detect Roman Fortified Sites: The Case Study of Blad Talh (Tunisia Section)

Round 1

Reviewer 1 Report

Thanks for the nice paper and overview,

 

a few improvements and questions:

 

Line 311 and 313, 316 , 483 still show spelling errors. 

 

In Figure 4 B the label is missing. 

 

How did the authors calculate the kappa coefficient?

 

Has the algorithm been applied to another image dataset with archaeological sites or at sites where sites are suspected?

Author Response

Dear Review 1,

The authors greatly appreciate your in-depth and thorough evaluation of the manuscript. Your valuable comments improve our manuscript a lot. We incorporated almost all your comments and express our deepest gratitude for the valuable contribution you made to improve this manuscript..All modifications to the manuscript are highlighted in red font for easy access. The point-to-point responses to the reviewer comments are listed below. Your comments are in black color and us are responses are in blue .

Author Response File: Author Response.pdf

Reviewer 2 Report

The submitted paper is a very interesting one and the author has correctly explained the topic they want to analyse. Optical and SAR data are used correctly and the method is good enough. Results are presented according to the objectives of the study. The results are well discussed and they are able to draw a fruitful conclusion. However, there are some minor fixes/changes that need to be addressed.

 ·         Is there any advantage to using polarization data in site selection? It needs to be added and discussed in the discussion section

·         Delete lines from 90-100

·         Line 172, change (Figure 1) to [Figure 1(b)]

·         In Table 1, in the spatial resolution column, change all values 2,14 to 2.14

 

·         Line 196, add NIR before red, green and blue.

Author Response

Dear Review 2,

The authors greatly appreciate your in-depth and thorough evaluation of the manuscript. Your valuable comments improve our manuscript a lot. We incorporated almost all your comments and express our deepest gratitude for the valuable contribution you made to improve this manuscript.All modifications to the manuscript are highlighted in red font for easy access. The point-to-point responses to the reviewer comments are listed below. Your comments are in black color and us are responses are in blue .

Author Response File: Author Response.pdf

Reviewer 3 Report

The manuscript aims  to evaluate potential offered by the combined use of Remote sensing data and machine learning to  discover fortified sites in arid areas..  This study is based on using remote sensing data combined with machine  learning algorithms . The proposed approach is not completely clear, some steps have to be better explained:

Abstract:
      1- Authors are requested to add the information how the P1B and SAR have been processed    in Google’s Earth Engine?

2-      Authors are requested to add more results.

Introduction:

1-      Authors did not elaborate novelty of this work hence there is room to make this as a useful contribution.

2-      Page 2 line 63, support this sentence with proper references such as

doi:10.1088/1742-6596/1973/1/012149, https://doi.org/10.1007/s12518-021-00363-6

3-      Page 2 line 78, please (begun to search) not begun To search.

4-      Page 2 line 78, support this sentence with proper references such as doi.org/10.1155/2021/5580286,

5-      The authors  specify the concept of archaeological sites and provide the appropriate references to deal with it.

6-      The authors should add more literature about the studies conducted on archaeological sites.  

7-      At the end of the introduction section, the novelty of this work should be provided.

Methodology

1-      Clarity of methods section, as written, the methods section is not repeatable. Much effort is needed to more specific in describing steps taken and cite all algorithms used.

2-      The authors failed to clearly explain, some steps have to be better explained in the methodology sections.

 

3-      The authors refer that the results evaluated from a comparison with field survey data in the abstract section. But they do not clarify how they use survey data in methodology section.

 

4-      Data collection and methods are not well captured hence suffering from many shortcomings.

5-      Authors are requested to add more justification for the data used.

.

Results

1-      The results needs more information to offer a scientific product to be used as new approach.

 

 

Discussion

1-      Authors are requested to add more discussion

1-       Overall, the manuscript needs English editing to by revise with better readability

 

Conclusion

1-      Authors are requested to add future recommendations for the study area

 

 

Author Response

Dear Review 3,

The authors greatly appreciate your in-depth and thorough evaluation of the manuscript. Your valuable comments improve our manuscript a lot. We incorporated almost all your comments and express our deepest gratitude for the valuable contribution you made to improve this manuscript..All modifications to the manuscript are highlighted in red font for easy access. The point-to-point responses to the reviewer comments are listed below. Your comments are in black color and us are responses are in blue .

Author Response File: Author Response.pdf