Efficiency Study of Combined UAS Photogrammetry and Terrestrial LiDAR in 3D Modeling for Maintenance and Management of Fill Dams

Round 1

Reviewer 1 Report

The provided manuscript is of very low scientific quality with significant problems in language editing and terminology. I will address the main issues chronologically:

line 17 - What do you mean by "recall of the point cloud-based 3D models were compared"? This should be better explained. 

line 51 - should be "The TLS and the point cloud from UAS photogrammetry processed on SfM principles..."

Table 1. - What is the meaning of this symbol ※? What is the reason for two different values for length 640 (625) and Dam crest width 6 (6.4)? (): is an unusual use of parenthesis.

Figure 2. - The first mention of SCPs shown in figure 2 is on line 195 - add the explanation of individual types of points in the figure description. SCP1 is showed in the figure twice.

Figure 3. - Registration - "sustem" instead of "system"

line 173 - Images were manually captured? On line 171 you state you used automatic flight with 85% overlap. There were two variants of camera network configuration during two flights? One with nadir images and one with camera under 60°? Please, explain.

line 174 - If there were two different flight configurations, what was the GSD in them? Or was it the same? How did you compute the 1.7 cm/pixel value? Based on flight height or based on the distance from the side of the fill dam (for the 60° images)?

line 177 - What was the overlap between images? 80%, 85%, or 75%? Make it clear, because you mention various values in different parts of the text.

line 181 - "In photometry" should be "In photogrammetry"

line 191 - TLS Survey - how many TLS scans were created? Only two? Or at every GCP and CP? Explain or add figure with positions of the scans.

line 2O2 - Section 2.6 - How did you register the TLS scans? Based on CPs and GCPs? You should provide statistics from registration.

line 211 - replace every mention of "3D numerical model" with "point cloud"

line 212 - "Pix4Dmapper automatically extracted the junctions centered on the image unit features and matched them with multiple adjacent images." Did you mean "Pix4Dmapper automatically oriented the UAS images based on features used as tie points"?

line 214 - "shooting position" should be "exterior orientation parameters"

line 214 - "mutual expression elements" should be "relative orientation parameters"

line 216 - WGS84 is not an image coordinate system...

line 217 - again "mutual expression elements"... "relative orientation parameters"

line 219 - I don't understand this sentence: "Thereafter, self-calibration based on the absolute coordinates was performed by reconstructing the 3D model, and a 3D numerical model was established through dense matching." Self-calibration can not be performed by reconstructing a 3D model... Self-calibration is a part of bundle adjustment during SfM reconstruction of the scene. The detailed 3D model can be reconstructed in the next phase, after image alignment.

line 223 - what are "station mark" and "backsight" coordinates? Wrong terminology...

line 224 - "collimating the backsight"? I don't understand.

line 245 - "CP is based on the planar Cartesian coordinates"? You did not measure the CPs in 3D? Only in XY plane?

Table 3,4,5 - unnecessary data presented - the absolute values of GCP and CP coordinates are not interesting, it would be much better to provide differences between the GNSS and photogrammetric measurements.

line 251 - "The model was composed of 220,220,220 point clouds." What do the number 220 represent? Is it the number of points in the point cloud? Then you should write "The model was generated based on 220 million points in the point cloud." A 3D model consists of triangles (or faces), not points...

line 266 - The errors presented are based on CPs that were signalized by artificial targets on plain surfaces without vegetation - So the statement about the effect of vegetation is false. You should analyze the effect of vegetation based on the cross-sections in figure 4. Looking at the scalebar in the figure, it seems that there were errors of around 1 m relative to the "floor plan" or project...

line 276 - I do not understand the recall rate. What does it mean, how did you calculate it, and why? Does it mean that you left 98.47% of points for further analysis after cleaning the original point cloud from vegetation? How did you decide which parts of vegetation to remove? Was it done manually or automatically with classification tools?

line 287 - You are referring to a point SCP0, but there is no such point in figure 7.

line 295 - "recall of the total area of the fill dam as 74.59%" - please explain better, this is confusing. Maybe there is a problem with the translation...

line 309 - How was the aerotriangulation (AT) done? Which software did you use? How did you manage to combine the point clouds from TLS and UAS-SfM? Only based on GCPs and CPs? The aerotriangulation makes sense only for photogrammetric projects with images. If there are only point clouds available, you can not use AT, only ICP. But with such significant changes caused by vegetation, the ICP could lead to inaccurate results.

Figure 9. - There should be an overview image with the location of details a) and b) - it is not clear where are these areas located on the dam.

Table 8 - Missing units for "Heaving Volume"?

Table 9,10 - Missing units for "Settlement Volume" and "Heaving Volume"?

line 388 - Confusing conclusions - the RMSE presented is based on CPs and does not represent the overall accuracy of UAS photogrammetry - the authors did not evaluate the errors caused by vegetation in point clouds.

I will address the conclusions according to the numbering from line 420:

1 - The conclusion is obvious - we can always obtain more reliable data when using photogrammetry and TLS in vegetation-free seasons.

2 - The "advanced 3D numerical model" is not sufficiently described. It is not clear how the two different sets of data (TLS and UAS-SfM) were combined and how exactly the ICP algorithm was used and with what results. Statistics from point cloud alignment are missing.

3 - Throughout the whole manuscript there are significant terminological problems. Recall rate and reproducibility are confused. It is not clear how the authors get the numbers presented and based on what criteria.

4 - The values of volumes calculated were probably not corrected by the influence of vegetation and thus are not reliable.

Author Response

Hello. I am the author of the paper you reviewed.

Thank you for your interest in my paper and for your high-quality review.

Thanks to you, it was a time to confirm that my ability to write a thesis has improved.

Author Response File: Author Response.pdf

Reviewer 2 Report

The article provides an intriguing study on 3D Modeling for the Maintenance and Management of Fill Dams using TLS and UAV data. Despite the problem having been addressed for several years, there is still a lack of effective solutions, making this work a valuable contribution. The article effectively highlights the novel aspects of the study.

However, there are a few language errors present in the paper. It would be beneficial to conduct a thorough check to ensure the text is polished and free of any mistakes.

Author Response

Hello. I am the author of the paper you reviewed.

Thanks to you, it was a time to confirm that my ability to write a thesis has improved.

Author Response File: Author Response.docx

Reviewer 3 Report

Dear authors,

The paper seems to be good prepared disussing a combination between UAS and TLS in dam study. The data are well collected and proccessed. Below are a few suggestions:

1.    Title seems OK. But, is this research mainly about the accuracy of UAS and TLS in dam study? Should make this clearer in the title. For example: The Accuracy of UAS Photogrammetry and Terrestrial LiDAR Combined in 3D Modeling for Fill Dams: A Case Study of OO reservoir in South Korea

2.    Abstract should be improved. The research method should be briefly introduced, namely how did you combine UAS with TLS in this research?

3.    Lines 13-25 are all the research results, which should be more concise. And the conclusion is only a verification of the possibility of application?

4.    In the paper, there are several places you used “examine the possibility”, this seems to be a very simple work. It will be better change into “evaluate the accuracy”

5.    Line 93, you listed many references [28,43,44,49,55–57] or TLS [22,41,54,58–60], can you briefly introduce them, especially the studies out of Korea? What method they used and how was their accuracy?

6.    Line 43, start a new paragraph to state the use of UAS and TLS.

7.    Figure 1. Because this journal is open for all scholars, please also highlight it in the location map to let international readers know where is this dam in Korea.

8.    The workflow 2.3 should eb moved before the section 2.1.

9.    The results/discussion are detailed and clear enough.

Good luck

 

 

 

Author Response

Hello. I am the author of the paper you reviewed.

Thank you for your interest in my paper and for your high-quality review.

Thanks to you, it was a time to confirm that my ability to write a thesis has improved.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

I have no further comments.