Review Reports

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

My main comments are in the attached PDF... here are some over arching comments:

- Too few CPs and a varying number of CPs used in RMSE calcs so incomparable. - NDVI comparison seems linked to threshold choice and feels unscientific (or perhaps I am missing something).
- Conclusion and Abstract imply better results than the study showed.

Comments for author File: Comments.pdf

Comments on the Quality of English Language

I have point out many small language issues, my general comments are below:

- Language inconsistent... switch between US and UK English (e.g. "meters", "normalized", "summarized" and then "colour", "normalised" and "optimised")
- some of your word choices need review (see PDF).

Author Response

Dear Reviewer,

We sincerely thank you for your thorough and insightful review of our manuscript. The revisions you proposed significantly enhanced the clarity and rigour of the article, and we sincerely appreciate the time and effort you invested in providing such thoughtful feedback.

Your primary concern is the number of verification points used to assess geometric accuracy. While we acknowledge the validity of your comments, we would like to explain our limitations in this regard. The terrain in the study area features nearly inaccessible sections and is densely covered with vegetation, making it impossible to measure signalised control or verification points. Additionally, no suitable natural points were available for use.

The primary focus of this work was to evaluate radiometric accuracy, as the geometric accuracy of products generated from UAS imagery with RTK is already well-documented. For the geometric accuracy assessment using a single GCP during bundle adjustment, we utilised 16 verification points, which we believe is a reasonable number given the challenging conditions. When performing bundle adjustment with 12 GCPs, the geometric accuracy is expected to improve or at least match the accuracy achieved with a single GCP. Consequently, the accuracy figure derived from using only five verification points was intended more as a control to check for potential blunders than as a definitive measure of accuracy.

We have incorporated this explanation into the manuscript's results section and hope it addresses your concerns satisfactorily.

Thank you for pointing this out, as it allowed us to clarify our methodology and findings.

Responses to the remaining comments and suggestions have been addressed in the "Comments" section of the provided PDF document. Additionally, the edits reflecting these changes can be reviewed in the Word document, where the "Track Changes" feature has been enabled to highlight modifications.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Summary Comments:

The study assesses the precision and reliability of the DJI Phantom 4 Multispectral (P4MS) UAS using Agisoft Metashape for geometric and radiometric correction in shrubland mapping. It evaluates the differences in the number and configuration of Ground Control Points (GCPs) for geometric correction and assesses the accuracy of using reference panels, an Incident Light Sensor (ILS), and both for radiometric correction. Reflection rates and NDVI are calculated, and the accuracy is evaluated using the Micasense Altum AL0 camera. However, the study's novelty is not apparent, and there are issues with the expression in some parts of the experimental method. The experimental design can further discuss the sample size and the reproducibility of experimental conditions. It is recommended to optimize the paragraph arrangement in the article's structure to more clearly highlight the main points of the research. For instance, the methods section is lengthy; adding summary paragraphs could help readers understand the logic of each experimental step. Linguistically, some paragraphs could be more concise and clear to enhance readability. Formatting-wise, the numbering is not uniform, and it is suggested to enhance the clarity of images.

Specific Comments:

l Lines 65-67: In the method introduction, it is mentioned that "several radiometric calibration methods exist." It is recommended to cite more literature and highlight the advantages and breakthroughs of the current research methods.

l Lines 156-162: An explanation of the reason for measuring wind speed or its role in the study could be included.

l Line 273-278: The numbering is not clear enough; it is suggested to unify the numbering throughout the text.

l Line 325-326: " Considering these factors, radiometric calibration converts raw imagery digital numbers to sensor irradiance and then to surface reflectance values." Please confirm if there is an issue with this statement's expression.

l Line 377: The heading numbering is not coherent.

l Lines 435-443: The experimental design can further discuss the sample size and the reproducibility of experimental conditions. It is suggested to combine Table 5 and Table 6 to make the comparison of results more intuitive, and at the same time, include table numbers in the language description for easy reference.

l Lines 447-448: It is recommended to describe this method in the main text of section 2.4, not just in the figure.

l Lines 609-612: Figure 19 should indicate the y-axis and use easily distinguishable colors to represent different cameras.

Comments on the Quality of English Language

No comment

Author Response

R#2: Lines 65-67: In the method introduction, it is mentioned that "several radiometric calibration methods exist." It is recommended to cite more literature and highlight the advantages and breakthroughs of the current research methods. 

Authors: The authors added more complete information to the introduction about radiometric calibration methods, including the advantages of the current research methods. These changes can be found in Lines 68-89: 

 “There are several radiometric calibration methods available,[16], [17] being the two most widely used the empirical line method  [14], [18], and the method that uses  incident light sensor (ILS) data [19]. Both have advantages and disadvantages [20]. The empirical line method converts raw digital numbers (DNs) to reflectance values based on the relationship between known reflectance values and their corresponding DNs in the image. Some studies have utilised two calibration targets with different grey levels [21], [22], [23], while others employed multiple targets to enhance calibration accuracy [24]. While this method is simple to implement, it assumes the surfaces (reference panels) have Lambertian properties (reflectance is not affected by illumination angle), a linear relationship between raw DNs (although some adopted nonlinear models that align with physical principles, establishing regression equations based on these models [25], [26], [27], [28]), true reflectance values [29], and that light conditions are the same throughout the flight. When the assumptions are not met, data quality and accuracy can be compromised. The ILS is integrated with the UAS and measures sunlight intensity in real-time, and allows adjusting the captured data to the changing light conditions. However, the possible lack of sensor calibration, the flight direction and the orientation of the ILS can affect the reflectance values due to light scattering in the atmosphere, and previous studies indicate a tendency of ILS to underestimate reflectance values [30].” 

R#2: Lines 156-162: An explanation of the reason for measuring wind speed or its role in the study could be included. 

The authors recognize that this information was missing. The importance of the wind speed and direction was added in Line 176-177:  

“Wind speed can impact on image quality, disrupting image overlap, affecting UAS stability, and introducing variations is flight altitude.” 

R#2: Line 273-278: The numbering is not clear enough; it is suggested to unify the numbering throughout the text. 

The authors changed the format to bullet points to add clarity. Line 296-302: 

• Without GCPs; 

• One GCP: five tests were done using one GCP, selecting the ones in the corners (GCPs 1, 3, 6, 8) and the one nearest the centre (GCP 12); 

• Five GCPs: one from each extreme of the study area plus the one nearest the centre (GCPs 1, 3, 6, 8 and 12); 

• Twelve GCPs; 

R#2: Line 325-326: " Considering these factors, radiometric calibration converts raw imagery digital numbers to sensor irradiance and then to surface reflectance values." Please confirm if there is an issue with this statement's expression. 

The authors recognize the sentence wasn't correctly formulated. The term “irradiance” was replaced with the intended term “radiance”. Line 353: 

“Considering these factors, radiometric calibration converts raw imagery digital numbers to sensor irradiance and then to surface reflectance values.” 

R#2: Line 377: The heading numbering is not coherent. 

The authors have updated the number throughout the text accordingly. Line 403, 404, 478, 871, 872 and 966. 

“3. Implementation and Results 

3.1 Geometric Accuracy Assessment” 

“3.2. Radiometric Accuracy Assessment” 

“4. Discussion 

4.2 Radiometric Accuracy Assessment” 

“4.1 Geometric Accuracy Assessment” 

“5. Conclusions” 

R#2: Lines 435-443: The experimental design can further discuss the sample size and the reproducibility of experimental conditions. It is suggested to combine Table 5 and Table 6 to make the comparison of results more intuitive, and at the same time, include table numbers in the language description for easy reference. 

The authors merged Tables 5 and 6 and added the n used for the statistical analysis. A detailed description on the number of VPs used (n) is in Lines 302-308. 

R#2: Lines 447-448: It is recommended to describe this method in the main text of section 2.4, not just in the figure. 

The authors opted for having a section “4. Implementation and Results”, instead of the common separation between Implementation/Methodology and Results to optimise the readability and easiness of interpretation of the manuscript. The authors recognise this is not the most common layout of scientific manuscripts, but given the nature of the content, it was found to be the best way to proceed.

R#2: Lines 609-612: Figure 19 should indicate the y-axis and use easily distinguishable colours to represent different cameras. 

The authors have updated Figure 19 accordingly and added more information for easier interpretation. Line 626. 

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

Good work on the improvements. It is nearly there... just a few things to fix (see comments in the PDF).

Comments for author File: Comments.pdf

Author Response

Dear Reviewer 1,

We sincerely appreciate your time and effort in reviewing our work. The answers to the comments are given in the .pdf attached, and the changes were done to the newly revised .docx.

With kind regards, 

The authors 

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have addressed my concerns. I have no further comments.

Author Response

Dear Reviewer #2,

We appreciate your time and effort in revising our work.

With best regards,

The authors