High-Speed Magnetic Surveying for Unexploded Ordnance Using UAV Systems

Round 1

Reviewer 1 Report

I strongly support the main goal of this investigation: effective and optimal utilization of the ROV magnetic survey.

However, this study is not free from a number of shortcomings.

(1) List of References should be significantly extended. Several references must be added in the Introduction. For instance,

Eppelbaum, L.V. and Mishne, A.R., 2011. Unmanned Airborne Magnetic and VLF investigations: Effective Geophysical Methodology of the Near Future. Positioning, 2, No. 3, 112-133.

B. Tezkan, J. B. Stoll, R. Bergers and H. Crofbach, “Unmanned Aircraft System Proves Itself as a Geophysical Measuring Platform for Aeromagnetic Surveys,” First Break, Vol. 29, No. 4, 2011, pp. 103-105.

A. Salem, K. Lei, C. Green, J. D. Fairhead and G. Stanley, “Removal of Cultural Noise from High-Resolution Aeromagnetic Data Using a Two Stage Equivalent Source Approach,” Exploration Geophysics, Vol. 41, No. 2, 2010, pp. 163-169.

(2) Do you calculate the influence of rugged topography? 

(3) Do you calculate the influence of oblique magnetization?

(4) Do you have some results of 3D magnetic field modeling of some typical cases?

(5) Do you have a comparison of magnetic surveys at different levels?  

Author Response

Dear reviewer,

Thank you kindly for these comments. Please find my answers below:

(1) List of References should be significantly extended. Several references must be added in the Introduction. For instance,

Eppelbaum, L.V. and Mishne, A.R., 2011. Unmanned Airborne Magnetic and VLF investigations: Effective Geophysical Methodology of the Near Future. Positioning, 2, No. 3, 112-133.

B. Tezkan, J. B. Stoll, R. Bergers and H. Crofbach, “Unmanned Aircraft System Proves Itself as a Geophysical Measuring Platform for Aeromagnetic Surveys,” First Break, Vol. 29, No. 4, 2011, pp. 103-105.

A. Salem, K. Lei, C. Green, J. D. Fairhead and G. Stanley, “Removal of Cultural Noise from High-Resolution Aeromagnetic Data Using a Two Stage Equivalent Source Approach,” Exploration Geophysics, Vol. 41, No. 2, 2010, pp. 163-169.

Great recommendation, I am always happy to expand on the references. The wordning in the introduction has now been broadened slightly to include general drone magentic surveying, as opposed to just being focused on drone magnetic UXO surveys. I have included the first two manuscripts that you suggested; you are absolutely right - they are highly relevant. I did not find the third manuscript to be a particularly obvious fit for the topic, so I did not include that. I did, however, add several other references instead (for a total of 11 additional references).

(2) Do you calculate the influence of rugged topography?
Not here, as the topography of the survey location was a completely flat beach comprised of sand and silt.

(3) Do you calculate the influence of oblique magnetization?
No, we refrain from making a distinction. the inversions demonstrated are thus not bound by any enforced direction, i.e., the magnetic moment is free to vary in both magnitude and direction given the model chosen for the inversion (this is not explained in the text, but instead in the references provided, as per the citation/referencing policy of the journal).

(4) Do you have some results of 3D magnetic field modeling of some typical cases?

I am a little bit unsure as to what 3D magnetic field modelling refers to in this regard, but I will attempt to answer regardless. We do not generate nor use any 3D shell/body models (since this would be well beyond the scope of our study). Our inversions are performed directly on the 3D point cloud of the magnetic data. The resultant source position as a single point in 3D space, and the equivalent dipole moment a 3D vector .

(5) Do you have a comparison of magnetic surveys at different levels?

Not in this study. Due to military exercises ongoing at the survey site, our time to collect data was unfortunately quite limited. We hope and aim to revisit this sometime in the future, if possible.

Reviewer 2 Report

The authors describe a novel configuration of quantum scalar magnetometers slung from a UAV. They test the system on deactivated examples of ordnance and test metal items. The new configuration displays excellent sensitivity and the options available for incorporating quasi-gradient information is also novel. The paper is well-written, well-described (given the patent application) and the experiments and results are convincing. The paper describes how various issues such as location, timing and orientation are adjusted for, along with external interference and the issue of sufficient sampling rates.

I commend the authors on their work and hope it leads to useful applications in real-world situations.

I have only two comments:

Figure 15 (b): The unit in (a) is a gradient so nT/m but the units in (b) are along track differences. In other figures, ATD they are given in nT/m from the scalar measurements in nT. To be consistent should the units here be in (nT/m)/m?

Figure 17: A similar question to above. In order to be consistent should the units be nT/m. ATD can be nT if they are sample to sample differences directly but here a '~1m' distance is mentioned, so could the units be nT/m?

Author Response

Dear reviewer,

Thank you for these comments, and for your kind words. I also hope that this will see real world application. If the things currently set in motion hold, this should happen no later than 2023 (but we shall see).

In relation to you comments:

Figure 15 (b): The unit in (a) is a gradient so nT/m but the units in (b) are along track differences. In other figures, ATD they are given in nT/m from the scalar measurements in nT. To be consistent should the units here be in (nT/m)/m?

Great catch - The ATD units were mislabelled originally back in Figure 12 (where they should be nT, instead of nT/m). The nT/m obtained when differencing across two sensors solely comes from the 1m seperation between the sensors. I have now corrected this.

Figure 17: A similar question to above. In order to be consistent should the units be nT/m. ATD can be nT if they are sample to sample differences directly but here a '~1m' distance is mentioned, so could the units be nT/m?

This is also an excellent point. I actually struggled a bit with the decision on how to present this figure clearly. We used ~1m since the mean is around 0.97m, but can vary quite a bit due to oscillations, wind etc. To hopefully make things more clear, I have now updated the figure such that the units are listed as nT/~m.

Reviewer 3 Report

Line 5 It is recommended to expand the abbreviation UXO earlier in the abstract in line 2
Lines 79-87 It is recommended to give more information on the on-board GNSS receiver. The following characteristics are worth to be mentioned: number of received channels and navigation signals (L1/L2/L5, etc.); positioning rate; positioning accuracy.

General comments
1. It is unclear from the paper which type of a UAV was used, a multi-rotor or a fixed-wing one.
2. There is almost no information on the flight planning procedure. Does it include preliminary aerial survey for getting the digital elevation model for considering the relief for flight planning. It might be not requred for plane topografphy. But what in case of hilly terrain?

Author Response

Dear Reviewer,
Thank you for this valuable input. I have adhered to all of your comments as outlined below:

Line 5 It is recommended to expand the abbreviation UXO earlier in the abstract in line 2

Good catch. This has now been done.

Lines 79-87 It is recommended to give more information on the on-board GNSS receiver. The following characteristics are worth to be mentioned: number of received channels and navigation signals (L1/L2/L5, etc.); positioning rate; positioning accuracy.

Excellent point. I have now added the relevant specifications for both the UAV and sensor positioning systems.

General comments
1. It is unclear from the paper which type of a UAV was used, a multi-rotor or a fixed-wing one.

Good catch. I have now clarified that we used a quadcopter.

2. There is almost no information on the flight planning procedure. Does it include preliminary aerial survey for getting the digital elevation model for considering the relief for flight planning. It might be not requred for plane topografphy. But what in case of hilly terrain?

Great catch. For this specific survey we did not generate/import a DEM (a quick LiDAR scan completed previously confirmed the surveyed area to be almost completely flat, with ~5cm variation in height across it) - but we would certainly generate such a model (using either LiDAR or photogrammetry) had the terrain been more challenging.
I have now stated the flight planning software (UgCS), and the possibility of introducing DEMs for terrain draping.