Research on a Measurement Method for the Ocean Wave Field Based on Stereo Vision
Round 1
Reviewer 1 Report
Paper title: Research on measurement method of ocean wave field based on stereo vision
Authors: Hanyu Sun , Guoqing Wu , Xueliang Wang , Tao Zhang , Pu Zhang , Wei Chen , Quanhua Zhu
Journal: Applied sciences, MDPI
General comment
The paper reports on the possibility to recover the wave field based on a stereo vision methodology, making use of a gridding SIFT GPU methodology.
The topic is interesting and has an engineering relevance if its applicability is proved.
However the manuscript reveals serious flaws and it must be strongly improved, including being rewritten by a native English speaker. For this reason I suggest that the paper is rejected but I encourage the authors to resubmit a new improved version of the paper, according (but not limiting) to some suggestions provided below.
Suggested items:
English must be strongly improved, it often makes the paper not readable due to the several grammar, punctuation, and construction errors.
There are many acronyms that are not properly clarified in the text (e.g. RBR line 23, DLT line 47, SIFT line 104, DoG, LoG line 109, and so on)
There are many symbols that are not properly explained (k, s, line 112, symbols in eq. 4 and following, and so on in most of the equations)
Many concepts are not explained (Moravec, line 70, Hough transformation, line 217, and so on)
The measurement units are missing almost everywhere.
Some images are not commented at all (e.g. Figs from 3-7)
The parameters of the sea surface (a, b, c q, f) need to be discussed and specified.
Author Response
Dear reviewers,
Thank you for your help and patience.
1、According to your proposal, English has been improved.
2、According to your proposal, acronyms have been properly clarified in the text .
3、According to your proposal,symbols have been properly explained in the text.
4、According to your proposal,concepts have been explainedin the text.
5、According to your proposal,the measurement units have been added in the text.
6、According to your proposal, images have been commented in the text.
7、The parameters of the sea surface (a, b, c q, f) have been discussed and specified in the text.
Reviewer 2 Report
Dear Authors,
the article entitled: Research on Measurement Method of Ocean Wave Field Based on Stereo Vision presents a complete method based on stereo vision wave measurement, using gridding siftGPU to achieve fast matching of large images, the running time of the whole process algorithm is within 6s, the number of feature matching pairs is guaranteed to be more than 20,000. It is very important because stereo vision technology can realize non-contact and mobile measurement, but the timeliness and adaptability are poor.
All chapters (abstract, introduction, materials and methods, results, discussion, as well as conclusions) are well described and they do not raise any doubts. In terms of the literature review is sufficient (32 positions), all of which are papers from recognised scientific journals, such as: Ocean Engineering, Remote Sensing, Sensors, and others. Moreover, I would like to point out that the papers cited are related to the subject of this article (stereo vision, gridding siftGPU, wave height, period and wave direction). However, in the publication make the following changes:
· I propose to extend the literature in first paragraph of the introduction, related to the importance of ocean waves in oceanography that are on the Journal Citations Reports (JCR) list such as for example:
- Makar, A. Determination of the Minimum Safe Distance between a USV and a Hydro-Engineering Structure in a Restricted Water Region Sounding. Energies 2022, 15, 2441.
- Specht, C.; Lewicka, O.; Specht, M.; Dąbrowski, P.; Burdziakowski, P. Methodology for Carrying out Measurements of the Tombolo Geomorphic Landform Using Unmanned Aerial and Surface Vehicles near Sopot Pier, Poland. J. Mar. Sci. Eng. 2020, 8, 384.
- Stateczny, A.; Burdziakowski, P.; Najdecka, K.; Domagalska-Stateczna, B. Accuracy of Trajectory Tracking Based on Nonlinear Guidance Logic for Hydrographic Unmanned Surface Vessels. Sensors 2020, 20, 832.
· In the introduction, it is worth summarising what the individual chapters of the article contain.
· It is neccessary to describe results in Figures 10-12.
· The article is not at the highest editorial level. It would be necessary to improve, e.g. errors and typos.
· Please write sentences impersonally
To sum up, after taking into account the above amendments (minor revision), I suppose that this review is suitable for publication in Applied Sciences.
Author Response
Dear reviewers,
Thank you for your help and patience.
1、According to your proposal, the literature in first paragraph of the introduction, related to the importance of ocean waves in oceanography that are on the Journal Citations Reports (JCR) list has been extended.
2、According to your proposal, the individual chapters of the article contain has been summarised in the introduction.
3、 According to your proposal,the results in Figures 10-12 has been described in the text.
4、 According to your proposal,English has been improved in the text.
Reviewer 3 Report
In this work the authors proposed a stereo vision method suitable for wave measurement, this approach uses gridding siftGPU to achieve fast match of large images, the running time of the whole process algorithm is within 6s, the number of feature matching pairs is guaranteed to be more than 20,000; and uses the least square method and sea surface wave surface theory to realize the calculation of the sea surface base level without control points. Compared with the RBR wave height meter, the measurement error of wave height is less than 10% and the period error is less than 0.5s, the wave direction error is less than 1°. In general, I general the manuscript is well written and deserves to be published. Here some minor issues which I encourage the authors to consider:
1 – I can understand the whole approach proposed by the authors however, why the authors chosen for a feature matching approach. Direct methods shush as based on Jacobians or Hessians could be more efficient in terms of hardware requirements/processing time.
2 – Maybe other descriptors such as RBRIEF (Rotated BRIEF), ORB or BRISK could be more efficient and does not require GPU acceleration.
3 – The authors estimate the 3D information by a calibration algorithm however, there are several different techniques which could be used (planar holography, linear triangulation, etc.) in any case, more discussion about this issue could be a nice complement of the current manuscript.
Author Response
Dear reviewers,
Thank you for your help and suggestion
1、 Direct methods based on Jacobians or Hessians can obtain the distance between the camera and the measured object,not the actual height of the measured object. Due to the change of wave tide level,the paper calculates the zero point coordinate of wave height every 10min to ensure wave height measurement accuracy.
2、Other descriptors such as RBRIEF (Rotated BRIEF), ORB or BRISK have been tested.Because the camera resolution is too large, the computer memory will make mistake during batch processing.Descriptors such as RBRIEF (Rotated BRIEF)、ORB 、 BRISK 、surf or sift have similar matching time based on siftGPU.
3、 According to your proposal,I will learn more about technology in the next work
Thank you for your help and patient.
Round 2
Reviewer 1 Report
I can't see a real improvement with respect to the previous version of the manuscript. Methods are not adequately described and conclusions are not properly discussed.
Author Response
Dear reviewer,
Thank you very much for your e-mail informing us the editorial decision on our manuscript (No: 1782662). We would like to express our gratitude to you for the time and effort spent in processing and improving our paper. The constructive comments from you are very appreciated which are extremely helpful for the improvement of our paper.
This is to confirm that the paper has been duly revised in accordance with the comments. Therefore, we would like to resubmit for a possible publication in the Robotics and Automation. Attached are the detailed responses to you and the duly revised manuscript.
Once again, sincere thanks for your time and effort in further processing our revised paper. Looking forward to hearing from you regarding its disposition.
Please see the attachment
Sincerely yours,
Hanyu Sun *, Guoqing Wu *, Xueliang Wang, Tao Zhang, Pu Zhang, Wei Chen, Quanhua Zhu
E-mails: sunhanyu@cssrc.com.cn, 15852797071@139.com, wangxl@ccsrc.com.cn,
zhangtao@cssrc.com.cn, zhang_7775@163.com, willenchan@126.com,
zqh666xx@163.com
E-mails: sunhanyu@cssrc.com.cn
Author Response File: 
Author Response.pdf
