Coastal Aquaculture Extraction Using GF-3 Fully Polarimetric SAR Imagery: A Framework Integrating UNet++ with Marker-Controlled Watershed Segmentation

Round 1

Reviewer 1 Report

This manuscript presents an approach for coastal aquaculture pond mapping using GF-3 fully polarimetric SAR imagery and developed a novel framework. The article is potentially interesting. But I would like to suggest the following list of revisions/comments to the manuscript to improve the quality of the manuscript further:

 

1.       In Intro, much research about mapping coastal aquaculture ponds using optical remote sensing was reviewed, but more related work was not mentioned. For example, studies on mariculture mapping were highlighted in lines 59-65, but there are numerous works about coastal aquaculture pond extraction using Landsat, Sentinel-2, and others that were not referenced. Please revise this part.

2.       In lines 93-96, the limitations of Sentinel 1 are related to its spatial resolution, and the size, and density of the aquaculture pond. For larger ponds, the boundaries can be accurate extraction. So, I think the advantages of GF-3 need to be highlighted in another scope.

3.       The caption of Figure 4 is not clear. And, the subplots of Figure 4 are hard to tell the difference.

4.       The Innovation of this manuscript is accurate mapping, so I think some detailed result examples in Figure.7 need to be shown.

5.       In Line 169, please change “breeding area” to “aquaculture area”.

6.       Please let us know the time cost of model training and classification.

7.       The method in this study is very promising for accurately mapping smaller-scale aquaculture ponds. The superiority of larger-size aquaculture ponds might be not outstanding compared to smaller size. So, can some differences be given about the superiority of the method on different aquaculture sizes?

 

8.       Suggest to polish the English writing by native-speaker.

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

The ms is overall interesting. It should be improved - to some extent - in some sentences that should be ameliorated both in terms of English usage and in clarity of exposition. I have some concerns that mainly deal with the features and a lack of physical explanation of the behavior expected for the features over the selected areas.

p3 r 102 "when using these traditional shallow learning methods", what do you mean for "shallow learning"?

p5 r 192 "containing multi-look" is not correct, may be you implemented multi-looking...

p6 the SI index is not clear. In particular, this Reviewer would like to know under which SI values the two classes are separate enough. Is there a threshold? The authors wrote that 0.8<SI<1.5 call for "an authentic feature". It is not clear to me what does it mean? Does it mean that the separation is fair enough?

In table 1 I do not understand the features "SE", "SE_I" and "SE_P". The Shannon entropy should be the polarimetric entropy derived from the eigenvalue decomposition, isn't it? What about SE_I and SE_P. It is interesting to note that SE and SE_I are the features calling for the best separability. hence, it is important to clarify their meaning, you can think to add a ref to each row of Table I to allow the reader to easily understand the meaning of the features.
Probably the Authors refer to "J. Morio, P. Refregier, F. Goudail, P. C. Dubois-Fernandez and X. Dupuis, "A Characterization of Shannon Entropy and Bhattacharyya Measure of Contrast in Polarimetric and Interferometric SAR Image," in Proceedings of the IEEE, vol. 97, no. 6, pp. 1097-1108, June 2009, doi: 10.1109/JPROC.2009.2017107." but it should be made clearer
Please, note that a list of polSAR features to be used for marine applications is also included in F. Nunziata, A.Gambardella, M.Migliaccio, “A unitary Mueller-based view of polarimetric SAR oil slick observation,” International Journal of Remote Sensing, vol.33, no.20, pp.6403- 6425, 2012. You can refer to it to provide additional info about the selected features
Another reference that may be useful is this one: A. Buono, C.R. De Macedo, F. Nunziata, D. Velotto and M. Migliaccio, “ Analysis on the effects of SAR imaging parameters and environmental conditions on the standard deviation of the co-polarized phase difference measured over sea surface,” Remote Sensing - MDPI, vol. 11. no. 1, pp. 18-33, 2018.

I would also appreciate the Authors show the VV-polarized image of the area of interest and the SNR achieved in the water-covered part of the aquaculture and on the dike. I would like to know whether the signal is corrupted by noise or not in the low backscatter area.  The SE and SE_I features should both shown as images since they are the features exhibiting the lagers separation among
classes. Hence, I would like to see that images. please, use also a colorbar to appreciate the range of values.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

The manuscript Article has been well-revised, and no further comments left.

Reviewer 2 Report

The ms can now be accepted for publication

thanks for incorporating the requested changes