A Partition-Based Detection of Urban Villages Using High-Resolution Remote Sensing Imagery in Guangzhou, China

Round 1

Reviewer 1 Report

The paper presents a method for the detection of Urban Villages in Guangzhou. In general, the work is well related, but there are parts that are not clear and their contribution is not highlighted. Although statistics are shown, no images of results are shown. Some issues are:

L24: Meanwhile.... that. It must be delated.

L30: would? Conclusions can not be conditional

Guangzhou is not a keyword

Whci is the difference between UV from others IUS?

L70. This paragraph mus be before. It is important to understand UV.

L47. This paragraph explains the quality of results of other methods but not the methods themselves. The comparison between results should be discussed, unless it is intended to detract from the overall quality of other work. The methods of other authors should be explained here.

Section 2.2.2 Why are the four districts divided into five case studies? Justify here and not later

Tablke 2. Add equations and/or references.

Section 2.4.1 is not clear

Section 2.4.2 is poor explained

L221. This paragraph is not well explained. It is not clear why 2x2hm grid, or why more than two UV and non-UV

Must indicate number of samples, percentage of training, validation and testing, processing equipment, processing time and classification

The difference between RF and SVM in Producer Accuracy in D and E is very striking in Table 4. This should be mentioned and justified.

L238: The explanation of values in table 5 is not clear.

L308-320: It is not clear

The conslusion is poor. The use of the conditional always causes doubts about the work of the authors. (e.g. "...lands could be..."???). Here, authros should give clear and concise answers about your work

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Dear authors,

Your study on “A Partition-Based Detection of Urban Villages Using High-Resolution Remote Sensing Imagery in Guangzhou, China” deals with the complexity to map informal urban areas in China (urban villages) using an OBIA approach. The paper is interesting and shows the potential of OBIA for urban village mapping but also the limitations, e.g., in terms of transferability across different urban zones in Guangzhou. However, the paper misses at present to explain several very important methodological aspects; besides, the paper would also require in other parts a major revision. Please improve the paper about the following aspects:

• Abstract: the abstract is not easy to understand for a reader without reading your paper. Thus, the abstract needs substantial improvement.
• Abstract: what do you mean by “spatially and temporally differentiated land urbanization” what are land urbanizations? I guess you mean simply urbanization?
• Abstract: Line 18: how was this division done – based on what?
• Abstract Line 20: better performance of what and compared to what?
• Abstract Line 24: the concept of zones is not clear - and in general, it is not clear what is classified here?
• Abstract Line 28: Please revise here and at other parts “land urbanizations”?!
• Throughout the paper, please add before every reference a space. For example, line 38 change to “and human progress[ 1,2].” -> change to “and human progress [1,2].”
• Introduction in general: it seems that “high-quality urbanization development” (I would change to planned urban developed) seems to be the main urbanization strategy. However, UVs offer affordable housing to low-income groups, which is important.
• Paragraph between line 60-70 maybe also add some work on building extraction,g., a new study: https://www.mdpi.com/2072-4292/12/10/1574/htm
• Section 75-81: urban villages also exist in other countries, it is part of urban expansion. You seem to suggest that is is only a phenomenon in China. I know this phenomenon for my own work in Cairo, Accra but also in Indian cities.
• Study area and data: please add a bit of information about the morphological characteristics of UVs.
• Study area and data: Line 100 - Can these images be freely obtained?! And please add information about the image characteristics, bands, spatial resolution etc.
• Line 112: “As demonstrated by other studies,” -> please add references – it is not clear which studies you are referring to.
• Line 113 etc: please be aware that the densities and building height in informal / slum areas differ a lot between cities and countries. For example, in several Indian cities or also in Cairo, multi-story structures in informal areas are common.
• Section 2.2.2. it is not clear how the kernel density analysis and the partitioning was done. Please explain this in this section in more detail.
• Line 149 and at other places: it is not clear what you mean by “and normal construction lands”.
• Line 154: What was the window size used for the GLCM calculation and how was this optimized?
• Table 1: Indicators – it is more common to refer to image features.

 

• Table 1: Why did you not use other GLCM features?

 

• Table 2: Can Tab. 1 and 2 be combined? There is some redundancy.

 

• Line 184: Why was ESP (estimate scale parameter) was not used? See e.g.: https://doi.org/10.1080/13658810903174803

 

• Section 2.4.2 and 2.4.3:  Please add here information about the training and the testing data distribution. It is not clear how training and testing data were selected, e.g., the number whether they are independent etc.

 

• Section 3.2: Did you map only one UV per zone or several?! The extent of your classification is not clear.

 

• Figure 6: Is this the result of your classification or a reference map?

 

• Figure 8: In this figure, you show Kappa and not the accuracy!

 

• Line 368: “First of all, the current kernel-density-based partition on the 2015 construction land…” -> I do not understand this statement – please revise.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

The authors have significantly improved the manuscript. The answers have been clear and well justified. I recommend its acceptance and publication