A New Automated Ship Wake Detector for Small and Go-Fast Ships in Sentinel-1 Imagery

Round 1

Reviewer 1 Report (Previous Reviewer 2)

The authors describe a novel technique to detect a ship's wake on Sentinel orbital imagery. Advantages of the article: 1) The authors propose a new technique to catch a ship's wake. 2) The authors implemented the technique on Sentinel's images. 3) The authors thoroughly measured the performance of the technique. Disadvantages of the article: 1) The authors do not compare the performance of the proposed technique with others known from the literature. It is tough to assess the solution's novelty without such a comparison.

Author Response

Based on the review provided, the following improvements have been made. The English language in the paper has been checked and certain modifications have been made to improve it. Following, the specific observation provided by the reviewer is addressed, it should be noted that the response to the observation is now addressed in the paper as well.

Contrary to previous state-of-the-art, the images analysed covered a widely extended area of ocean. The authors deem this factor to be determinant in not allowing a fair comparison to be made amongst different techniques presented in previous papers.

We thank the reviewer for the time taken in providing us useful feedback.

Reviewer 2 Report (New Reviewer)

The paper proposed an automatic method for ship wake detection in SAR images through the preprocessing, bright and dark line detection and error checking. Here are some suggestions for the authors.

1. The contributions of this paper in Introduction section should be stressed.

2. "Ship detection is a vastly researched topic due to the wide range of possible applications, from maritime security to traffic planning and illegal activity monitoring such as drug trafficking and illegal fishing." This sentence is repeated in the introduction so the rewritten is needed to make the introduction more logical.

3. Figure 1 and its corresponding description of different wakes should be more clearer. Now the turbulent wake and narrow V wake are confused from the image.

4. For the detection problem, detection  time is another important evaluation for the algorithms. Thus, the paper should illustrate or discuss about the time. 

Author Response

Based on the review provided, the following improvements have been made. The English language in the paper has been checked and certain modifications have been made to improve it. Following the specific observations provided by the reviewer are addressed, it should be noted that the responses to these observations are now addressed in the paper as well.

The contributions of this paper to the field of study have been deepened in the introduction section providing an overall view of the techniques utilised and the novelty that these bring to the field.

The image has been improved by indicating more clearly the various components of the wake. Finally the efficiency and low computational value of the techniques used that allows for a potential real-time analysis of the area under investigation have been underlined.

For the second point, the authors are seeking clarification from the reviewer about exactly what reorganisation or change in the introduction they need to bring. The reviewer states "this sentence is repeated in the introduction" but the authors could not find where this happens.

The authors thank the reviewer for the time taken in providing us with the useful feedback.

Reviewer 3 Report (New Reviewer)

This paper par excellence presents an automatic wake detection method for small and fast ships. This paper is well written, the method is sound, and the results are convincing. Before publication, I have some minnor suggestions for the authors:

1. Line 162 "In the context of this paper, for go-fast ships the authors mean ships visible in the image only by means of their wakes." In my experience, if the ship is not visible, the wake may also be difficult to see. Therefore, I suggest that the authors highlight the ability of the proposed method to detect the wake of such ships in the experimental part.

2. I suggest that the authors, in their future works, test the ability of the proposed method under different sea states, and improve their method in high sea states.

3. Perhaps, "A New Automatic Ship Wake Detection for Small and Go-fast Ships" may be a better paper title.

4. The introduction lacks in reviews of recent works in SAR ship detection. For example：1) A Two-Stage Method for Ship Detection Using PolSAR Image. IEEE TGRS, 2022. DOI: 10.1109/TGRS.2022.3216532. This work shows an effective method for small ship detection. 2) GPU-Oriented Designs of Constant False Alarm Rate Detectors for Fast Target Detection in Radar Images. IEEE TGRS, 2022. DOI: 10.1109/TGRS.2022.3188151. This work reports a very fast  GPU -CFAR algorithm for SAR ship detection.

Author Response

Based on the review provided, the following improvements have been made. The English language in the paper has been checked and certain modifications have been made to improve it. Following, the specific observations provided by the reviewer are addressed, it should be noted that the responses to these observations are now addressed in the paper as well.

The application of the wake detector is of particular interest in situations where there is a good likelihood that the wakes are visible even if the corresponding vessels are not. A detailed analysis of how the RCS of a wake could change under different circumstances was beyond the purpose of this paper but constitutes an area of future study for the authors. Similarly to what stated previously, in the future we aim to consider different sea state conditions and how they affect the RCS of the wakes and their persistency.

Moreover, the title has been changed to A New Automated Ship Wake Detector for Small and Go-Fast Ships in Sentinel-1 Imagery.

Finally, one of the suggested studies, namely A Two-Stage Method for Ship Detection Using PolSAR Image. IEEE TGRS, 2022, has been added to the state-of-the-art discussion. The other paper was analysed by the authors but was deemed not relevant to the discussion at hand.

The authors would like to thank the reviewer for the time taken in providing us useful feedback.

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 presents an automatic ship detection method in Sentinel-1 imagery by exclusively identifying their wakes. Experiments have verified the effectiveness of the method. This paper is interesting. I only have some minor concerns.

1) The writing of the paper should be improved. There are lots of long sentences that are hard to fully understand.

2) There are some typos. For example, line 64 ‘dependant’. Besides, the reference [14] and [15] seem placed at wrong locations in lines 63 and 67.

3) Figure 3 (the diagram of the proposed algorithm) is not clear enough in the paper.

4) It is suggested to provide some example figures to demonstrate how the true positives are obtained by the detected lines.

Reviewer 2 Report

The article describes an algorithm for vessels detection on SAR images.  Advantages of the article: 1) A new algorithm is introduced. 2) A clear description of the method used. 3) Laboratory performance steps are given. Disadvantages of the article: 1) It is unclear what novel components the algorithm introduces. 2) Performance of the algorithm is not compared with any other known from the literature. 3) Scientific soundness can be easily improved by more references to literature younger than 5 years.

Reviewer 3 Report

see attached file

Comments for author File: Comments.pdf