Development and Testing of Octree-Based Intra-Voxel Statistical Inference to Enable Real-Time Geotechnical Monitoring of Large-Scale Underground Spaces with Mobile Laser Scanning Data

Comment

Response

1

The reviewed article deals with the development of a new method of statistical testing of data obtained from mobile laser scanning of mining objects in order to detect their changes in real time. For this purpose, the authors developed a method based on statistical testing of differences in scanned point clouds in voxels and compare it with the state-of-the-art M3C2 method. I consider the development of this new method to be a fundamental contribution of the reviewed article, relevant to the topic of mobile laser scanning.

The authors also prepared a rich overview of the current state of the problem. The references sufficiently describe the topic treated in the article. I consider the conclusions resulting from the presented research to be correct. The key results are summarized in Fig. 7, 9 and in Table 3, which compares the computational complexity of the new method and the M3C2 method.

NA

2

I recommend publishing the article after small formal modifications - unification of the font size in the text, the Gigabyte unit has the abbreviation GB, not Gb.

We appreciate the comment and changed the style of two paragraphs to the correct MDPI text style as they were incorrectly formatted as figure captions.

We changed the abbreviation to GB.

Comment

Response

1

In this study, a novel octree-based computational framework for intra-voxel statistical inference change detection and deformation analysis is presented. The study is comprehensive and interesting in terms of its subject. However, some revisions are required.

2

In the Abstract section, the results should be briefly mentioned without going into details. Discussion of the results and recommendations for the future are not included in this section.

We appreciate the advice and removed details from results, and discussion-like content, and also removed future recommendations.

3

The titles of sections 3.2 and 3.3 are the same.

We appreciate the comment – we changed the name to the correct title for Section 3.3: Change Detection Accuracy Tests

4

I suggest that you present your results in sections 3.2 and 3.3 as tables. It will be helpful to the readers.

We appreciate the comment and added Table 2 and Table 3 to summarize the results in a more structured format.

5

The number of the conclusion title should be 4.

We appreciate the comment and changed the title number to 4.

6

Conclusion Section can be written more concisely.

We appreciate the comment and reduced the length of the conclusion by about 20% by removing redundancies and improving conciseness.

Comment

Response

1

The work is useful work. It can be published in your journal if the following corrections are made.

2

Fahle et al. [20] showed that multi-epoch MLS data could detect geotechnical hazards while achieving data quality with uncertainty on the millimeter-to- centimeter level. I did not come across) I would be glad if you explain.

Our previous work can be found here: Analysis of SLAM-Based Lidar Data Quality Metrics for Geotechnical Underground Monitoring | SpringerLink

3

What is the difference between the work you have done and other work? What kind of awareness has he added to science? Make a comparison.

We made changes in the Introduction, Octree-based deformation analysis and Change Detection to highlight how are method differs from previous work. The updated text includes the following:

·         “Previous voxel-based change detection work focused on relatively large and discrete changes in urban environments and often used high-accuracy lidar sensors. In contrast to these methods, …”

·         “Unlike conventional change detection techniques employed in underground applications, our method needs to produce a robust binary change classification…”

We also revised our conclusion section to be more concise and showcase the scientific advances we demonstrate in our work:

·         “In comparison to the state-of-the-art M3C2-based approach, our method…”

·         “Additional benefits of our approach include a more automated workflow and the ability to store other data within its efficient octree-data structure”

4

The literature should be expanded a little more.

We added two additional references to the introduction, including a recently published review on laser scanning for underground applications by Singh et. al. (2023).

As stated by other Reviewer 1, we believe the review of current literature is otherwise comprehensive and relevant. If Reviewer 3 would like to suggest specific works that we have not covered, we are more than happy to consider any additional relevant research.

5

What is the calibration value of your device before starting the measurement?

We were not entirely certain of what the Reviewer 3 refers to by “calibration value”, we interpreted it to refer to any setup process required for the Emesent Hovermap MLS.

The Hovermap MLS used in our study performs an automated calibration process when connected to power. After this calibration process (approx. 30 s - 1 min) the user can start a scan. The scan is preceded by a short scan initiation (< 10 s). Both processes are fully automated and do not require any user input.

We added an explanation in Section 2 – Field Data Acquisition.

6

In these systems working with SLAM, measurement speed and route are important. It would be helpful if you put Trajectory in the article.

We appreciate the comment and added details for the FDS 1 and FDS 2 dataset scanner orientation relative to the targets in Section 2 – Field Data Acquisition. We also added the average scan trajectory speed for the FDS 3 dataset.