Discrimination of Steel Coatings with Different Degradation Levels by Near-Infrared (NIR) Spectroscopy and Deep Learning
Round 1
Reviewer 1 Report
I do have several comments and suggestions:
► The SEM cross-section image of the coated sample should be given in the Materials and Data Collection section.
► The Description in Table 1 should be supported with SEM images of the coated sample after the four-level coating conditions.
Author Response
We thank the reviewer for his/her constructive comments.
The reviewer has a valid point, We understand that the reviewer’s suggestion on the microstructure analysis using SEM is important. The initial experimental design included SEM test for microstructure information and EIS test for electrochemical information. Unfortunately, our laboratory is inadequate and has no equipment to perform such tests. We also have limited access to other schools due to covid 19. We will definitely consider this as one of our future directions of research.
Reviewer 2 Report
The manuscript has been well written. So, it can be accepted in current form.
Author Response
We thank the reviewer for his/her approval.
Reviewer 3 Report
This work concerns the investigation of Steel Coating Condition by Near-Infrared (NIR) Spectroscopy and Improved 1D Convolutional Neural Network. The quality of paper is unacceptable – even though structure is clear and overall impression in good, after careful reading language has to be corrected, and no interesting results can be found. I found some mistakes/flaws and both major and minor comments have to be taken into account, in order for this work to be published in Coatings:
MAJOR Comments:
a) First and foremost problem of this work is quality of presented NIR results and absolutely no difference in mean spectra given for four different coating condition systems, as well as unjustified attributions of “troughs” to some vibrations:
· With no differences between given spectra, neither novelty of this work nor formation of combined NIR+CNN model can be stated
· Authors provided numerous post-processing methods for spectra, but In my opinion baseline correction was missing as a very important part that should highlight the presence of very doubtful bands described by Authors,
· Authors described presence of peaks and “troughs” – I have never experienced attributions to “troughs”. Moreover literature sources are missing in this part, which only makes results more doubtful,
· Authors should provide spectra in wavenumbers [ cm-1 ] instead of wavelengths [nm],
b) Language has to be corrected (I suggest the help of native speaker) in terms of numerous grammar, spelling and punctuation mistakes i.e.:
· 29th line 1st page – what Authors mean by “pose significant safety” ??
· 40th line 1st page – Authors should use past tense – “For example, Piehl et al. [6] presentED first…”
· In next sentence Authors should not use past tense – “..they were able to IDENTIFY AND QUANTIFY…”
· What Authors mean by “abrasion of coatings thickness” ?
MINOR Comments:
a) What kind of steel was used and is it the particular steel used for HZMB bridge ? – this information is missing,
b) What kind of softwares were used for post-processing of data, as well as preparation of CNN model ?
c) I do not agree with Authors statement that C, O, N and S are “relatively heavy atoms”.
Author Response
Please see the attachment.
Author Response File: 
Author Response.pdf
Reviewer 4 Report
The quality of the current form of the manuscript is not appropriate for publication.
Even I don't feel the paper is related to coating characterization as there is no SEM and XRD analysis. Simply the authors have executed data analysis using soft computing techniques.
However, I am suggesting some points below:
1. In the abstract the sentence "Near-infrared (NIR) spectroscopy is a rapid, low cost and non-destructive analytical technique that can be used in characterization the performance of steel coating" was written. So, what type of characterization you are doing? Explain it. Also, add the important results in the abstract.
2. Introduction is so specific. A critical review was missing. Better to add a literature review separately to discuss the research gap in assessment methods of Steel Coating.
3. Section 2 is inadequate.
Add the sample preparation methods.
Add the coating procedure
Add some real images of coatings
Also, add the material specifications with SEM, XRD, and microstructure images and discussion.
3. From where you got Table 3 data?
4. Provide some references for section 3.2.
5. Elaborate Figures 1a and b clearly.
6. Discussion in section 4 was need to be improved by considering the literature opinion.
7. Conclusion should be written point-wise.
8. Title should be Rewritten- Assessment of Steel Coating Condition by Near-Infrared (NIR) 2 Spectroscopy and Improved 1D Convolutional Neural Network
Think more about bold words. Remove NIR
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Round 2
Reviewer 3 Report
Corrections are accepted and paper can be published in the present form.
Reviewer 4 Report
The authors have made significant corrections. Now it can be accepted for publication.
