Next Article in Journal
Spatial Spillover Effect of Digital-Finance-Driven Technology Innovation Level Based on BP Neural Network
Next Article in Special Issue
Evaluation of Land Use Efficiency in Tehran’s Expansion between 1986 and 2021: Developing an Assessment Framework Using DEMATEL and Interpretive Structural Modeling Methods
Previous Article in Journal
Correction: Yang, J.; Yu, M. The Influence of Institutional Support on the Innovation Performance of New Ventures: The Mediating Mechanism of Entrepreneurial Orientation. Sustainability 2022, 14, 2212
Previous Article in Special Issue
A Comprehensive Review on Land Use/Land Cover (LULC) Change Modeling for Urban Development: Current Status and Future Prospects
 
 
Article
Peer-Review Record

Assessment of Machine Learning Methods for Urban Types Classification Using Integrated SAR and Optical Images in Nonthaburi, Thailand

Sustainability 2023, 15(2), 1051; https://doi.org/10.3390/su15021051
by Niang Sian Lun 1, Siddharth Chaudhary 2,* and Sarawut Ninsawat 1
Reviewer 1: Anonymous
Reviewer 2:
Sustainability 2023, 15(2), 1051; https://doi.org/10.3390/su15021051
Submission received: 11 October 2022 / Revised: 13 December 2022 / Accepted: 28 December 2022 / Published: 6 January 2023

Round 1

Reviewer 1 Report

The current manuscript describes a new way to classify urban land use by combination SAR and indicators such as NDVI, NDWI, and NDBI. The writing is generally good, and the reviewer had the following comments:

1) The introduction provided some discussions on using SAR to categorize land cover. However, it misses discussing the shortcoming of the current methodology. The purpose of the introduction is to identify current knowledge gaps. If no knowledge gap is identified, there is no need to have the new methodology proposed by the current study. In short, the introduction needs to identify the problem of current methodologies and how the current study solves such problems.

2) Section 2.2.2: Please identify the mode (SM, IW, EW, or WV) that the data is collected.

3) Figure 4: “Restairants” is a typo, “upto” should be two words, and “Shopping Centers” is using two different font sizes.

4) Figure 4: How do you classify buildings of 13-39 stories?

5) Figure 4: The reviewer is not sure about the situation in Thailand, but many SE Asian countries have mixed land uses in cities, which means that a building can have residential as well as commercial zoning together. How do you classify areas in such cases?

6) Section 2.4.1: Please provide a flow chart for the steps.

7) Line 207: Please delineate how the orbit file is obtained. The ESA website does not provide orbit files.

8) Section 2.5: Please delineate what data is used in calibration and validation. How did the authors extract training data from the eight zones depicted in Figure 2?

9) Lines 338, 339: Please use subscripts for equations.

10) Line 371: If gamma =1 gives the best results, then VVH = VV. Why not just use VV?

11) Lines 411, 416, 421: Shouldn’t they be subsection titles?

12) Section 3.6: What is the validated accuracy?

13) The logic behind the current study is that each land use has its special building pattern. This might not be true for older cities where the construction and city planning practices change over time. How do you apply this methodology to such places?

14) A lot of citation formats do not follow MDPI requirements. In-text citations must be numbers. Besides, it is wrong to use citations as the subject of a sentence. Please correct.

Author Response

Dear Reviewer,

Thank you for your valuable suggestions. Please find the attached document in which we have tried to address your technical comments and also made the changes in the manuscript.

Thanks

Author Response File: Author Response.pdf

Reviewer 2 Report

As combining SAR with optical satellites for urban classification is not new, authors need to provide a clear justification why this study was needed?

Define abbreviations before first use even in abstract

Be consistent with naming satellite datasets by using images only. Currently the paper uses images, pictures, and photos.

The manuscript needs to be further refined as there are many redundant and repeated sentences.

Most of the tables missing captions.

Some tables and figures included but not mentioned anywhere in the manuscript

One of the assessment shown in the results named as “F1 score”. But it is not discussed in the methodology how it is obtained and why needed?

This study classified the urban areas into three classes: residential, commercial, and other areas. However, by looking at figure 4 characterizing these three areas, it uses only building heights (40-4 floors) for residential, while it uses only building types (offices and shopping center) for commercial areas. This base of grouping the three classes needs to be explained otherwise it is not clear and not correct.

Double check the English especially the use of tense

Line 17: Nonthaburi. In which country?

Line 21: Remove “the” from “…the three classes…”

Lines 21-23: To avoid confusion, it is better to name the classification as residential buildings, commercial buildings, and others

Line 24: Remove “the” from “…the three-machine learning…”

Lines 33-34: These abbreviations are not defined beforehand

Keywords: Missing optical satellites

Lines 38-44: Add reference

Line 113: Add reference

Lines 121-122: Elaborate more on how the 8 zones were selected based on the three designates forms of the city. Something like: zone 1 represents - - - - form of the city and so on.

Lines 129-130: No need for the satellites history, but dates, paths and rows of the acquired images are needed here

Table 1: Change “period” to “Acquisition date”

Table 1: Provide exact date with days

Line 137: What is NOSTRA?

Line 141: It is not clear how authors were able to use google maps to obtain updated numbers of floors comparing to the old 2012 NOSTRA data?

Line 173: Elaborate more on these 16 parameters. What are they?

Line 174: You mean “imagery inputs” otherwise they are three data inputs

Line 187: How the 12 parameters mentioned would match with the 16 parameters mentioned in line 173?

Line 191: Avoid “such as”. List exactly what has been used in this study.

Line 230-235: Re-arrange these bullets by embedding them in the paragraphs of section 2.4.1

Lines 260-262: Who developed this ML algorithm? When? Does this mean you used this algorithm which was developed for Hangzhou City in this study current study?

Lines 262-315: These lines are literature review; they don’t belong this section.

Line 333: “respectively” of what?

Line 335: Avoid “for example”. List exactly what has been used in this study.

Line 141: “The model was developed …”. Which “model”? It is better to use same terminologies for easy connection when reading the manuscript.

Figure 5. Explain what “Ln Height” means?

Equations 3.1: double check its number

Equations 3.2 and 3.2: Should move up into methodology section

Figure 6: Indicate clearly the units of X and Y axis

Section 3.4: There are no results in this section. It can be moved to the literature or methodology sections.

 Section 3.5: Out of 27 lines of this section, only 2 lines contains results other line are repeated information.

Author Response

Dear Reviewer,

Thank you for your valuable suggestions. Please find the attached document in which we have tried to address your technical comments and also made the changes in the manuscript.

Thanks

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Thank you for providing a revised manuscript and answers to the comments. The reviewer still has the following comments:

1) The introduction is only minimally edited, which is not enough to highlight the knowledge gap. Why don't you include your answers to comments in the manuscript?

2) Please include answers to the comments (e.g., mode of data used) in the manuscript.

3) Out of curiosity, please identify the orbit file (in the manuscript or as a response to comments) from downloaded SAR data package. The reviewer used SAR data from Setinel-1 before but cannot find the orbit file.

Author Response

Dear reviewer,

Thank you for your valuable comments. Please find attached file in which authors have addressed your comments.

Thanks

Author Response File: Author Response.pdf

Reviewer 2 Report

It was difficult to track how the authors incorporated the reviewer’s comments. Stating “Other comments related to grammar/ spelling mistakes addressed in the manuscript” is not enough. Authors should address each point by their response and stating the location of their changes in the revised manuscript or state their reason why that particular point is not incorporated.  Nevertheless, the revised version has improved better than original one.

For more improvement of the paper, the following points are still need to be fulfilled:

Be consistent with naming satellite datasets by using images only. Currently the paper uses images, pictures, and photos.

Most of the tables missing captions.

Some tables and figures included but not mentioned anywhere in the manuscript

One of the assessment shown in the results named as “F1 score”. But it is not discussed in the methodology how it is obtained and why needed?

Table 1: Provide exact date with days

Line 141: It is not clear how authors were able to use google maps to obtain updated numbers of floors comparing to the old 2012 NOSTRA data?

Line 173: Elaborate more on these 16 parameters. What are they?

Line 230-235: For better connection with previous paragraphs, Re-arrange these bullets by embedding them in the paragraphs of section 2.4.1.

Line 335: Avoid “for example”. List exactly what has been used in this study.

Author Response

Dear reviewer,

Thank you for your valuable comments. Please find attached file in which authors have addressed your comments.

Thanks

Author Response File: Author Response.pdf

Round 3

Reviewer 1 Report

Thanks for providing the answers to my questions. There is no more questions.

Back to TopTop