Efficient Lightweight Surface Reconstruction Method from Rock-Mass Point Clouds
Round 1
Reviewer 1 Report
The goal of the research is very well set and presented, in the manuscript introduction. Also, the contributions are well described and explained.
The related work is extensively analyzed and well presented.
The methodology and results are well organized and presented with a sufficient level of detail.
Finally, the conclusion chapter summarizes correctly the achievements of the research. Also, the trade-offs, that authors had to make are presented along with possible future research.
Author Response
Thank you for your review and evaluation. I wish you every success in your work.
Reviewer 2 Report
Congratualtions to authors on very interesting article.
Author Response
Thank you for your review and evaluation. I wish you every success in your work.
Reviewer 3 Report
This article addresses this issue of extracting geometric structure of surface from point cloud. The method presented is logical and reasonable. Because this subject is still under extensive studies, this article could tentatively contribute to the development.
- This article uses the term “lightweight reconstruction”. This reviewer suggests to provide a brief definition. “Lightweight” itself is a sort of undefined, and usually is relative to a “standard”. A more direct term may be better.
- While seven datasets were used for the experiments, it is not clear why these seven are selected and what are the reason for the selection. These seven were taken from the same repository. Is the methodology developed specifically for the samples in this repository?
- Is there a scheme for the assignment of parameters? What are the sensitivities of the parameters to the final result?
- While the agreement of point clouds to the linear surface model formed may be useful, the RMSE could be mainly influenced by the either measurement noise or surface roughness. Meanwhile, the linear surface set selected may not be measured with this scheme. How these issues are addressed?
- The accuracy assessment is only performed with Rock1 and the man made model was used as the true reference. First of all, the manual operation could be manipulated and biased. Secondly, what is the guideline for this manual operation, and its uncertainty? Is it made by one person? And, only once?
- How the topological error is assessed? Why only Rock4 has topological error evaluation?
- What is "Rebustness text" in the title of 4.4?
Author Response
Thank you very much for your suggestions, which are very helpful to modify my article.Please see the attachment.
Author Response File: Author Response.docx
Reviewer 4 Report
In this paper, an efficient method is proposed for 3D modeling from point clouds. The motivation is clear, and the methods are clearly presented. However, I mainly have some comments about the experiments. 1. “Reconstruction for point clouds” usually refer to generating 3D point clouds or 3D shapes from multiple images. I would suggest that the authors could change the title to “surface reconstruction from point clouds” or “3D modeling from point clouds”. 2. In the experiments, there is no quantitative results showing the quality of reconstruction comparing with other baseline method when testing on Rock 3. Meanwhile, why using different baseline methods when testing on Rock 3 and 4. Since the test scenes are quite limited. It would be better that the author could provide a more comprehensive comparison to prove the superiority of the proposed solution. 3. Section 4.4, it should be robustness not “rebustness”. 4. In the experiments, only one dataset was used for testing the robustness. However, it is obvious that when the noise level increased, the coverage metric improved. It is hard to make a conclusion about the performance of the proposed solution under the increase of noise level. I would suggest that more datasets could be used for testing for proving the robustness. 5. Since many parameters are involved in the proposed method, it would be better that experiments on testing the sensitivities of some important parameters involved could be conducted.Author Response
Thank you very much for your suggestions, which are very helpful to modify my article.Please see the attachment.
Author Response File: Author Response.docx
Round 2
Reviewer 3 Report
This reviewer wishes to thank the authors for their diligent efforts in the revision. Glad that the suggestions may be constructive. There is no further suggestions from this reviewer.
Reviewer 4 Report
I don't have any further comments about this paper.