The InflateSAR Campaign: Developing Refugee Vessel Detection Capabilities with Polarimetric SAR

Round 1

Reviewer 1 Report

This is an interesting paper dealing with Refugee Vessel Detection Capabilities with Polarimetric SAR. However, the manuscript content is too lengthy, and the conclusion is blurring and missing. Shorten the whole manuscript to identify new important points more clearly, compared to the previous publications of authors (references 9 and 10). 

 

Other points to be revised are as follows:

1. According to the definition of loss tangent, equation (1) seems wrong. Check the equation precisely. 

tan delta = e’’/e’

 

2. The description of resolution and pixel size of each observation data should be provided. The resolution is a very important factor to distinguish objects. It is better to add the spatial resolution of the data in Table 2 as well as wavelength of each platform. 

 

3. The font size of legend and axes is too small to read in Fig. 4. In addition, the horizontal axis range of HV is different from those of HH and VV. Unify the range for intuitive comparison.

 

4. There is no unit (maybe in meter) in Fig. 5. Specify it meter or pixels.

 

5. What is AUC? Although it appeared in page 12/26 without explanation, but it suddenly appeared in Fig. 11. Define and explain the content and usefulness before using it.

6. The same is true for ROC. 

 

 

 

 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

 

The paper develops a method to detect vessels in distressed situations to mitigate the challenges faced by the influx of refugees who use these vessels. First, a dataset with realistic features is collected to train a suitable model. Then nine suitable detection systems are tested on the generated dataset and the results are analyzed.     The paper studies an extremely interesting problem with practical significance. The following comments should be addressed for the next round of reviews:     1. Since this paper is on a societally important problem, I recommend including a paragraph on the statistics of refugees and the vessels they use in the introduction section to highlight why this research is practically important.   2. The problem of vessel/ship detection using deep learning in the SAR domain has been studied in recent work from different angles of challenges. Consider:   a. Chang, Y.L., Anagaw, A., Chang, L., Wang, Y.C., Hsiao, C.Y. and Lee, W.H., 2019. Ship detection based on YOLOv2 for SAR imagery. Remote Sensing, 11(7), p.786.   b. Zhang, S., Wu, R., Xu, K., Wang, J. and Sun, W., 2019. R-CNN-based ship detection from high-resolution remote sensing imagery. Remote Sensing, 11(6), p.631.   c. Rostami, M., Kolouri, S., Eaton, E. and Kim, K., 2019. Deep transfer learning for few-shot SAR image classification. Remote Sensing, 11(11), p.1374.   d. Zhang, T. and Zhang, X., 2019. High-speed ship detection in SAR images based on a grid convolutional neural network. Remote Sensing, 11(10), p.1206.   e. Kang, M., Ji, K., Leng, X. and Lin, Z., 2017. Contextual region-based convolutional neural network with multilayer fusion for SAR ship detection. Remote Sensing, 9(8), p.860.   The above works should be discussed in the introduction section to provide a broader context to the reader.   3. Will the code be released to enable other researchers to reproduce the results?   4. In section 4, please also provide some directions for future work so researchers can benefit from this work for future explorations.   5. Please release the dataset on GitHub so others in the research community can benefit from this work.  

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Revised manuscript seems fine, except for 2 items:

Fig.3(a) is missing.

Fig. number is missing in line 402.

Author Response

Thank you very much, we corrected everything accordingly.

Reviewer 2 Report

The authors have addressed my concerns convincingly. My only suggestion is to spend some time and improve the presentation.

Author Response

Thank you very much for that suggestion. We reworked the whole manuscript and improved the quality of the language / grammar at several places. We further tried to improve the result presentation in the subsection “Detector Testing” and separated “Discussion” from “Conclusions”.