Novel UAV Flight Designs for Accuracy Optimization of Structure from Motion Data Products

Round 1

Reviewer 1 Report

Dear Authors,

Thank you for submitting an excellent manuscript.

Your paper is about an optimal path to build a DSM using low-cost UAVs. The paper is so flawless that it can be published without revision. However, some details need to be confirmed, and I think this will improve the quality of the manuscript.

1. According to the title, your article is mainly focused on 'accuracy optimization of structure'. The area used to prove the study is 225*253 m^2, and the elevation difference seems to be 15 m (p. 7, line 242). There are two specific questions here.

  - Why did you use DSM as a target for validation? For general use, wouldn't it be better to validate against a general target(objects like trees, buildings, or specific point objects) other than the DSM?

- When looking at the area and elevation difference, the area does not appear to have a severe elevation difference. Of course, it is not possible to know the irregularities of the area, but how can we say that the area is a proper area to show the validity of this study?

2. The authors are using HDM as a means of demonstrating their research achievements. Lidar DTM can be considered true values because that is created by credible organizations, so it can be considered credible data. However, I cannot confirm anything about the accuracy of that data. According to p.13 line 392 the resolution is expressed. But what about height accuracy? Is the accuracy high enough to support the success of this study?

Once again, thank you for sharing your wonderful research achievements through your manuscript.

Author Response

Please see the attachment. All new changes are highlighted in green in the new manuscript version.

Author Response File: Author Response.pdf

Reviewer 2 Report

The study focuses on a compelling and timely topic as UAVs are increasingly utilized in precision farming and cultural preservation. Developing strategies, like flight planning, to enhance the quality and reliability of products is undoubtedly a crucial aspect to explore.

The paper is well-written and provides a thorough analysis. I have a few comments to make.

The authors made the decision not to utilize any GCPs, despite having availability to them. This decision significantly impacted the outcome because they relied on a UAV system with only a basic navigation-grade GNSS sensor. However, this approach may be intriguing as there may be situations where ground control is not feasible. It would be helpful to provide further clarification on this aspect in the introduction section.

The variance in the results is mainly attributed to the ability to accurately estimate the camera's internal parameters due to the suggested acquisition planning. Since the authors calibrated the camera using the software wizard (chessboard approach), it would be insightful to assess whether utilizing this parameter set as initial values could decrease the variation resulting from the various strategies.

Some minor issues:

1.     Line 18: To my understanding, the most commonly used translation for the acronym UAV is "Unmanned Aerial Vehicles".

2.     Line 170: Traditional strategies such as nadiral and oblique have been around for a while, while POI is relatively new and less commonly used. It may be beneficial for non-expert readers to have a brief explanation of what POI is.

3.     Line 257: Including the mission year for UAV surveys in the report, even if mentioned later in the paper, could be helpful.

4.     Line 278-282 and Figure 4A: Why did not have introduced cross strip to increase robustness?

5.     Figure 4: Reporting GCP's position in all four images could also prove helpful.

6.     For POI, Spiral, and Loop flights, it is evident that the traditional concept of longitudinal and transversal overlaps does not apply. However, I needed to establish a time-lapse for taking pictures. What was this value? Was it consistent for all missions? Regarding spiral flights, selecting a constant time lapse results in a notable difference in image overlaps between the spiral's center and edge. Do you have any thoughts on this matter?

7.     Line 344: Please report on the quality of your GCPs measurements. Could you please specify which type of GCPs you used, whether they were artificial, natural, or something else? Thank you.

8.     Line 391: What is the quality of the Lidar DSM?

9.     Table 7: Please specify the unit of measurement for each given figure.

Author Response

Please see the attachment. All new changes are highlighted in green in the new manuscript version.

Author Response File: Author Response.pdf