A Review of Deep-Learning Methods for Change Detection in Multispectral Remote Sensing Images

Round 1

Reviewer 1 Report

The manuscript is well written and well structured and minor adjustments have to be done. In fact, would be better to consider to include also limits and potentialities of Deep Learning in mountain areas. In the manuscript it is not underlined this part considering that more then 30% of land are not in plains! If author will do it I will recommed this manuscript to be published on Remote Sensing.

In order to help you and including the Deep Learning algorithm application in geomorphological complex area here it is some useful reference to add in order to dicuss about this issue.

In particular showing limits and potentialities and asnwering the question what are the future application? What would be the role of the newest VHR satellite missions like Pelicans from Planet or Neo Pleiades of Airbus?

Here it is the reference mentioned:

- https://doi.org/10.3390/app13010390

- https://doi.org/10.1016/j.isprsjprs.2017.12.012

- https://doi.org/10.3390/rs13040584

The other pars are fine!

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Dear author,

you have presented a thorough review of deep learning methods applied onto freely available multispectral imagery to detect changes in scenes.

The paper is well organised, presenting a short summary of the methods lately published into the field.

No major comments in general.

line 10: ... development IS considered.

 

Thank you.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

This paper reviewed the deep learning based change detection of multispectral images. The introduced methods contain the main advances, to improve the manuscript, the following questions can be considered

1. What is the challenge of CD for multispectral image? Most of the methods  are designed for panchrmatic or RGB image, what is the extra information provided by more bands of multispectral image? More discussions can be made.

2. How to exploit the spectral information to improve CD? More descussion or suggestion can be made for future deep learning network.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf