Transforming 2D Radar Remote Sensor Information from a UAV into a 3D World-View

Round 1

Reviewer 1 Report

This manuscript aims at introducing a measurement system combining a UAV and low-cost radar sensor to reconstruct a 3D scenario. The topic is interesting and relevant. However, it can't be considered for publication until the following questions are clearly clarified.

1. English writing needs thoroughly revised by a native speaker. There are many typo errors in the text. E.g. in Line 480, page 13, ‘derivation’ should be ‘deviation’?
2. The authors are expected to review all of the formulas and explain the symbols in advance, E.g. formulas (6), (7), (8). Moreover, all abbreviations need to be explained on first use, such as synthetic-aperture radar (SAR) in Line 54, Page 2, and radar-cross-section (RCS) in Line 140, Page 4, etc.
3. In the Introduction section，the authors spend lots of text introducing the development of UAVs and the applications of various sensors. However, according to the title of this manuscript, it should focus on 3D reconstruction based on the data obtained by UAVs. I think it is necessary to introduce some UAVs for 3D surveying and mapping work, and compare them to illustrate the novelty and particularity of authors’ work. In other words, the motivation should be explained more clearly.
4. In Section 2.2, the configuration and parameters of the ARS-408 were introduced in detail, leading to it reads like a sensor manual handbook. It should be revised carefully. In addition, it mentioned that four datasets are recorded. What is the difference between the sensor data stored in the RPI and the data in the UAV flight controller? In other words, what is the difference between RPI sensors and UAV sensors in 2.3.1, only the temperature sensor is different? The authors should explain more clearly in the manuscript.
5. The proposed data processing method was inadequately described. It is better to start by providing the reader with a high-level figure of the whole problem.
6. In Section 2.3, the clustering algorithm was used many times. The authors are expected to give assessment of the complexity of the clustering algorithms and describe the efficiency of the proposed algorithm in real application.
7. In the Results section, it mentioned: “different flights show that an altitude of about 20 m and a flight speed of about 1 m/s is most suitable.” The settings of altitude and flight speed are for the specific scene of this article. How adaptable the algorithm is to different scenarios? For unknown scenarios, how to select suitable UAV flight parameters?
8. In the Results section, it also mentioned that the number of clusters was the same for the high and normal sensitivity modes (on line 436 page 12). Why? Can authors give more explanations?
9. Whether the barometer information and GNSS altitude are used at the same time? On the other hand, four flight tests were carried out in the experiments. What are the different purposes of the four flights? it is recommended to clarify them in detail.

Author Response

Please see our word doc.

Author Response File: Author Response.docx

Reviewer 2 Report

1) Combining the use of radar sensors within UAV platforms is an interesting and less investigated topic, as the authors correctly highlighted in the Introduction. To make such an aspect even more evident, I suggest authors to discuss how so far radars and UAVs have been mainly considered as separated technologies across different domains. For instance, radars have found a number of applications in UAV detection:

- "Detection and Classification of Multirotor Drones in Radar Sensor Networks: A Review", Sensors 2020

- "Review and Simulation of Counter-UAS Sensors for Unmanned Traffic Management", Sensors 2022

In my opinion, adding pointers to the existing literature would better motivate the interest in investigating a different and synergistic use of the two technologies.

2) It is stated that the average flight duration is around 15 minutes. Isn't it a too limited duration for a sensing campaign? Since the proposed system is specifically designed for geoscience applications, the total duration should be considered as a parameter of primary interest. Authors should better discuss this aspect and clarify how the proposed solution can be considered useful in practical scenarios.

3) Since there are some geometrical considerations behind the acquisition of the data clusters from the radar, as discussed in Sec. 2.2., I suggest authors to add a graphic representation of the acquisition process, so as to better contextualise the parameters reported in Tables 1-3.

4) In my opinion, the discussion in Sec. 2.3.5 about outlier detection should be extended with additional, possibly mathematical details describing how the filtering step is actually performed.

5) The same previous comment applies to Sec. 2.3.8. This section contains some geometrical processing that cannot be easily explained using text only. 

6) There are some typos scattered across the manuscript. Please proofread the whole manuscript carefully.

Author Response

Please see our word doc.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

The authors have responded to all my questions in detail, I have no questions now, so I recommend the paper for publication.

In addition，I also have a comment that might help to improve the quality of the paper.

Because the paper introduces a new measurement system and the eleven-stage data processing method seems complicated. It migh be better to draw a block diagram or flowchart of the overall scheme, so that readers can more quickly understand the main work the authors have made.

Reviewer 2 Report

The authors have correctly addressed all my comments.