Unsupervised Change Detection around Subways Based on SAR Combined Difference Images

Round 1

Reviewer 1 Report (Previous Reviewer 2)

Thanks very much for the authors' reply. There are no other problems on my side.

Author Response

Thank you very much for your positive comments on our revised manuscript. 

Reviewer 2 Report (Previous Reviewer 3)

The authors have resolved most of my concerns, However, the following reference should be discussed in the introduction, which are SAR data feature representation and application, e.g., 10.1109/LGRS.2022.3159179, 10.1109/LGRS.2022.3161931.

Author Response

Thank you very much for your positive comments on our revised manuscript! Thank you for your updated suggestions. We have supplemented the relevant analysis of the two literatures you recommended. Please find them in Line 71-79 of the newly modified manuscript.

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

This paper proposes a new weighting method for different Difference Images (Dis) when dealing with the change detection task. The proposed method is somewhat innovative and can achieve better results than existing methods. However, the original paper should be improved in order to be published, based on the reviewer’s suggested comments.

 There are some comments given as follows:

 1.      Please state more clearly in the abstract how innovative the proposed method is compared to existing methods.

2.      I wonder whether the introduction of CDI into change detection as an indicator is the original of this article. Please explain in the article.

 3.      The description on lines 101-115 is a bit confusing. It is suggested to clarify the proposed algorithm flow, and then describe the algorithm compared with it. The proposed algorithm flow cannot be seen clearly in the current description.

4.      In Equation (1), does D1 range from 0 to 1 ? Please explain.

5.      Section 2.1.2 only describes the implementation details of the algorithm. Please elaborate on why the weights between different DIs are designed in this way. Please elaborate on why the "energy" weights are named.

6.      In Section 4.2, 4.2.1 and 4.2.2 are suggested to be exchanged, that is, display single difference images first, and then display the combined difference images.

7.      In line 306, the mean of CoDI-DM is not high than values of CoDI-LEW, please confirm and describe more clearly.

8.      In Section 4.3, please confirm whether the reference image is generated with reference to the optical image. The description is inconsistent with the flow chart in Figure 1.

9.      I am confused about your meaning on lines 380-381, please describe it more clearly.

10.   I think the title of section 4.4 may be inappropriate and easy to cause confusion. This section compares the decision method of whether there is a change, not the whole algorithm. Please consider changing a more appropriate title.

11.   Please create hyperlinks to references in this article for easy reading.

12.   I think there are still some rooms for improving the language of this manuscript. Many grammatical and typographical problems should be revised. Please ask a native speaker for corrections in these cases. Also, please make sure that each subfigure is described and the subfigure description is correct.

Reviewer 2 Report

To obtain the more accurate information of surface changes around the subway, this study proposed a local energy weight method by combining difference images. Experimental results show the good performance of the proposed method. However, some issues should be addressed.

Major issues:

1) There are some problems in the structure of the paper. If the authors only choose the data of Wangfu Zhuang Station and Fangte Station of Jinan Rail Transit Line R1 for the experiment, then the data used in the experiment should be described in detail in the Section 2. What is the reason for choosing this set of data?

2) The introduction is too simple. This paper mainly discusses the change detection problem. In order to more comprehensively describe the related works, some other types of modes have also been proposed, such as spectral image and lidar image. It is suggested that the authors add the description of related work, such as

[1] Super-Resolution Mapping Based on Spatial-Spectral Correlation for Spectral Imagery [J]. IEEE Transactions on Geoscience and Remote Sensing, 2021, 59(3): 2256-2268.

[2] Adaptive Target Profile Acquiring Method for Photon Counting 3-D Imaging Lidar[J]. IEEE Photonics Journal, 2016, 8(6): 6805510.

In addition, the main contributions of this paper should also be given in detail in the introduction.

3) The experimental part needs to be improved. In order to further prove the universality of the proposed method, can the authors also conduct relevant experiments in other areas? It is also suggested that the authors add the evaluation index of operation time

Minor issues:

1) The variables of some formulas are not explained. For example, Pe in formula (15).

2) The format of the paper is not standard. Some words in the text are marked in red. In addition, there are some grammatical errors in the article, which need further careful proofreading.

Reviewer 3 Report

Comments on ‘Unsupervised change detection around subway based on SAR combined difference image’

This paper proposed a to use Local Energy Weight (LEW) method to establish Combined DIs and then use the FCM to achieve the change detection. Overall, the main contributions are not significant. All the individual DI generations are common to the readers. In terms of the LEW, it seems a simple combination of different DIS, therefore, the main contributions are weak. Some following concerns are:

1)     In the introduction, the motivation of fusing the multiple DIs should be discussed clearly. Moreover, the individual advantages of different DIs could be explained in detail.

2)     The descriptions of CDI are not clear. First of all, the value range of CDI, as well as the meanings of different values, need to be clarified. Second, what are the advantages and effects of CDI compared to DVDI and MVDI? What changes CDI can be described that cannot be characterized by DVDI and MVDI?

3)     In 2.1.2, the descriptions of the LEW are not clear. First of all, LEW is just a weighted sum of the different DIs. Are the value ranges of the three DIs consistent? Are the changes in their representation consistent? Second, what is the motivation for this simple summation?

4)     In the experiments, the figure 7 should give the individual DI image for comparison rather than the CD result map.