Frequency–Wavenumber Analysis of Deep Learning-based Super Resolution 3D GPR Images
Round 1
Reviewer 1 Report
The paper proposes a frequency-wavenumber (f-k) analysis incorporated with a deep learning-based SR network for unwanted noise reduction and electromagnetic wavefield decomposition.
In fact, the process is divided in two steps:
- 1 - super resolution (SR) images are generated by a deep residual channel attention network
- 2 - f-k analysis is applied on those images
The introduction is clear and the topic well exposed.
Part 2.
SR images enhancement is a very interesting topic and the authors should pay more attention to the part 2.
This part is really too short. What's new in this part? It seems to be a simple application of the method described in [39].
The scientific contribution must be exposed.
Part 3.
This part is well written, f-k analysis is well explained step by step. However, the scientific contribution must be better exposed. The reader is left with the impression that this is a course and not a scientific contribution.
Part 4.
The validation is well done.
Part 5. & 6.
Conclusion should be completed once the scientific contribution will be clearly exposed in part 2 and 3.
Author Response
The authors’ responses to the reviewers’ comments are attached. Each comment is addressed specifically below along with a brief description of how we have modified the manuscript to reflect the respective comments. We appreciate the reviewers’ comments and believe the reviewers’ comments have helped us make this manuscript a better document.
Please see the attachment.
Author Response File: 
Author Response.docx
Reviewer 2 Report
General comment
In this manuscript the authors propose the SR-GPR image enhancement network incorporated with the f-k analysis to improve GPR images and facilitate interpretation.
The work is interesting and useful to be applied as another multi-signal processing algorithm. However, in the conclusions, I find a brief reference to the fact that this proposed processing should also be combined with the others existing algorithms in order to obtain a much better final images. For example, the final images of figures 11, 14 and 15 would be greatly improved with an additional processing flow.
Comments and specific remarks
They are included in the attached PDF
Comments for author File: 
Comments.pdf
Author Response
The authors’ responses to the reviewers’ comments are attached. Each comment is addressed specifically below along with a brief description of how we have modified the manuscript to reflect the respective comments. We appreciate the reviewers’ comments and believe the reviewers’ comments have helped us make this manuscript a better document.
Please see the attachment.
Author Response File: Author Response.docx
Reviewer 3 Report
The article presents a novel filtration technique to enhance SR 3D GPR images. Advantages of the article: 1) The new filtration technique is proposed. 2) The effectiveness of the technique was verified in simulations. 3) The technique was implemented in real 3D GPR equipment. Disadvantages of the article: 1) It is not clear why filter (2) was chosen by the authors and how to establish the sigma parameter value? 2) The test results illustrated in Figure 8 are non-decisive. Low pass filtered signal is always better approximated by a spline. 3) Tests conducted by authors are assessed qualitatively only. 4) It is extremely hard to for the reader compare results presented with results of his own. 5) The authors do not compare their results with any other 3D GPR images enhancement known from the literature.
Author Response
The authors’ responses to the reviewers’ comments are attached. Each comment is addressed specifically below along with a brief description of how we have modified the manuscript to reflect the respective comments. We appreciate the reviewers’ comments and believe the reviewers’ comments have helped us make this manuscript a better document.
Please see the attachment.
Author Response File: Author Response.docx
Round 2
Reviewer 1 Report
The authors take in consideration the reviewer's comments. Their responses are clear. Thanks.
Reviewer 3 Report
Thanks for your answers. It clarified a lot to me.
