Lookup Table Approach for Radiometric Calibration of Miniaturized Multispectral Camera Mounted on an Unmanned Aerial Vehicle
Round 1
Reviewer 1 Report
I have a few comments before I can accept the manuscript for publication.
First, the introduction does not describe a knowledge gap that is being addressed by mentioning previous studies and how the proposed research improves upon other calibration methods.
Second, while developing the methods section, it was unclear how many images were taken for the LUT correction and how was this number of image decided.
Third, on lines 238 to 241 the authors mention that for the dark current correction they selected five pixel locations as being representative, how were these pixel positions selected?
Fourth, figures 4 and 5 need additional labels to make it clear which image corresponds to which band of the camera. The same comment applies for figure 7 and 8. Additionally, in Figure 7, the figure labels are no longer for the pixels but rather the bands of the camera? Please correct or clarify the change.
Fifth, figure 8 needs more work to correctly differentiate between all histograms (different alpha values perhaps).
Sixth, on lines 279-282 the authors mention that they selected the dark current correction based on gain values of 1x and 1ms integration times. I belive this choice needs a better explaination, in particular the chocie of the 1ms should be made clearer to the reader (i.e., perhaps becuase 1ms was a good compromise on the standard deviation that seemed to be in the middle of the other values).
Eighth, the axes and numbers of figures 12 and 13 need to be of higher resolution as it is hard to read them, additionally, the authors shoudl compare the standard deviation/RMSE of the images before and after the vignetting correction to provide a better measure of the errors.
Ninth, were the assumptions of linearity checked after performing the least squares regression? It is also not clear how many samples/points were used for the regression.
Tenth, the discussion section needs to be expanded to include future work and the implications of the study.
Author Response
Please see the attachement
Author Response File: 
Author Response.pdf
Reviewer 2 Report
The work presented is interesting considering the increase of available miniaturized camera both multispectral and hyperspectral available for UAV systems.
The text need to be improved from Chapter 3 to 5 and consequently in the discussion of results (chapter 6) and conclusions (chapter 7).
In Chapter 3:
Is well described but tables and figures are often not well described and linked in the text.
Figure 2 and Table 1 are not mentioned in the text therefore the reader has difficulties to understand what is shown in the figures.
Figure 3 is not mentioned in the text and it needs to be improved, possibly a scheme may be more effective.
Figure 4 is not mentioned in the text.
The experiment scheme description (figure 4) should be clarified in order to improve the understanding of operations sequence to create the LUT. In the text the experiment is described with a number of steps but the scheme (figure 4) shows that all procedures are performed in parallel.
Chapter 4 Calibration Result
the chapter needs to be reviewed and improved to permit the reader to follow the procedure and understand the results obtained which contribute to the construction of the LUT to correct the optical effect and noise.
Figure 5 is not mentioned in the text and it is not very clear, it shows 5 plots which should be referred to the 5 spectral channels but the caption is very poor and reference for X, Y axis are missed.
Figure 7-8-11-12-15 and table 2 are not mentioned in the text and the figure captions need a large improvement to describe the multiple plots presented within each figure, x, y axis reference most of the time missed.
Figure 14: in the figure caption is mentioned MicraedgeMX sensors which has never been mentioned before in the text, please give more details is not clear if it is part of MicaSense RedEdge-MX camera.
Chapter 5 Accuracy verification
Chapter 5 needs to be carefully reviewed and improved to show both the procedure to verify the correction approach.
Figure 17 and 18 are not mentioned in the text Moreover is not clear if the accuracy verification test has been made by the camera in a fix position (on a tripod?) or during a UAV flight.
Reflectance values for reference calibrated panels are shown in table 3 which again is not mentioned in the text. Please Authors describe in more details the link between the reference panels reflectance and radiances measured by the camera to make clear the approach used to retrieve values and to compare results obtained after corrections.
in figure 19 the plots seems to show reflectance levels (x,y axis reference not given).
Figure 19 shows 10 plots but the in the caption is not given any description of them, in particular is not shown the relation of the plots respect the geometry of the diagonal pixels selected for the text as described in figure 18.
5.3 Absolute accuracy
In this paragraph I suggest to improve the lines 415-415 by giving a more accurate description of x’ and Xi (The reality radiance of the radiation-415 target ?′ is the product of downward radiance and the reflectance of the radiation target). Is Important that the reader understand how the compared values are calculated.
Figure 22 is not mentioned in the text and I would like to suggest to describe the passage from reflectance to radiance to evaluate the absolute error showed in figure 23 which is not again mentioned in the text.
Chapter 6 and 7
Please improve the discussion and conclusion after the review of chapter 3-5, in particular the reader has difficulties to recall the LUT approach since the figures and results on chapter 3 to 5 are showing only some cases.
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Reviewer 3 Report
Please find attatched detailed rewiev in PDF file.
Comments for author File: Comments.pdf
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
The manuscript has been improved. However, there are a few details that need correction.
On L246 the Word cross-reference is missing " Error! Reference source not found.. ".
Figure 4 needs to be improved. It is now clearer that the corrections are applied in a sequential manner but at the bottom of the figure, the three correction factors seem to happen again before the final output.
Figure 5 has missing labels. The text describes five-pixel positions but there are only labels for three of them.
The authors have not yet answered if the least-squares regression complies with the linearity assumptions, specifically, was the homoscedasticity assumption verified?
Author Response
Please see the attachment.
Author Response File: 
Author Response.docx
Reviewer 2 Report
I thank the Authors for the improvement of the text, description of the procedure and for improving figures captions and their link to the text.
My only further request is to add in the text the reference to figure 11 and 12, paragraph 3.1.2.
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Reviewer 3 Report
Ok, paper has been corrected and experiment has been improved.
Still some typeing and editing issues.
line 223, 224 ?
246 reference not found
248 - reference to the producer website, producer name and country
246 - figure 66 ?
287 - 2x figure 7 ?
Figure 7 - almost not readable text size
463 - 2x figure 19 ?
Author Response
Please see the attachment.
Author Response File: Author Response.docx
