Research on an Underwater Target-Tracking Method Based on Zernike Moment Feature Matching

Round 1

Reviewer 1 Report

The subject of the paper is very interesting and nice method is proposed.

The research is generally well planned and performed, however some details in description is missing.

Therefore I propose soem remarks to be included for better clarification. 

Remarks:

- Introduction, L27-32; There were also a few approaches of using neural networks (like GRNN) for tracking of underwater targets - it is worth of mentioning

- L22 "better robustnes": it is not clear better than what?

- L75: Generally very nice description of paper genesis in introduciotn, however please supplement it with clear identification of the goal of the study and the research hypotesisi if any was stated

- L106: typically sonar gives the images. please provide information how do you get pointcloud, without measuring distance to echo

- L109: however it will probaly increase with time as in any dead reconing system. or not?

- L145: please provide information what is the difference and what is the relationship between target detection and feature extraction. It is not clear at this moment

- L171: can you please give simple example of what feature vector is to clarify this matter? And what are this 7 dimensions?

- L189: in which cordinate system?

- L191: please explain why size of the feature template is a circle

- L221: please explain why A = 1 and B = 1

- L302: please add information what was the depth of this iron block

- fig.10: bigger zoom would be better

All the best!

Generally fine. Small issues detected - please make double check

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Authors present an application of the PF using the Zernike moment matching. The filter is verified offline with data obtained from an experimental platform, an AUV. Three different sets of experiments to verify their algorithm are tested.

Despite the fact the data is processed offline, it is recommended to mention som aspects about the computer resources for data processing. This will complement the future work outlined in conclusions on the implementation of real-time object tracking experiments.

Comments for author File: Comments.pdf

Non.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Dear colleagues!

Thanks for you paper. It is very easy to read and very interesting approach for the AUV CV implementation. I did not see often implementation of Zernike moments to this matter. And your results are impressive.
Just a few moments I need to highlight.
The main advantages of Zernike moments include their compact representation, rotational invariance, and robustness to noise. However, they may not be as suitable for highly elongated or asymmetric shapes, as their orthogonal basis is better suited for circular and symmetric patterns. For more complex shapes, other moment-based methods or shape descriptors might be more appropriate. So may be it would be good if you tried to use a set of different targets for tracking. This is not a flaw in this article. Rather a suggestion to consider such options in the future.
As for this paper, for part concerned with the practical application, I did not see detailed description of tracking experiments in Maishan Reservoir, Zhoushan.
The readers might be interested what is the conditions you have in your pool: temperature, salinity (water is fresh or salted), currents presence. These details may affect on result.

Second question about algorithm implementation. You just showed algorithm and your AUV with target for tracking, but did not inform readers about practical implementation of algorithm for this AUV. Have you use Python or C++ for implementation, what is OS have manage AUV and sonar - these details might be also very interesting for the readers.

So more extended comments on this matter would be very appreciated. So my opinion - the paper is very interesting and important. I think you can easily get rid of these inconsequential flaws, and I looking forward to see your paper printed very soon.

With best wishes,

your reviewer.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf