Extraction of Forest Structural Parameters by the Comparison of Structure from Motion (SfM) and Backpack Laser Scanning (BLS) Point Clouds
Round 1
Reviewer 1 Report
This paper compared the accuracy and efficiency of SFM and BLS in extracting forest structural parameters. The content is comprehensive and has significance for guiding the extraction of forest structural parameters by SFM and BLS. Some revisions are recommended before further processing.
1 It is recommended that added a summary at the end of abstract.
2 Please add some explanation about the scanning style of “#” in line 163
3 An intact technology route is easier to understand the ms for readers.
4 The significance of this study needs to be further strengthened, especially in the abstract, introduction, discussion and conclusion.
5 Some language errors need to be carefully examined and modified, for example, line 164, “including” may be more suitable instead of “such as”.
Author Response
Please see the word file.
Author Response File: 
Author Response.docx
Reviewer 2 Report
This study uses TLS to assess the accuracy of forest structural parameters extraction with SfM and BLS points clouds. Overall, the paper is well-organized and straightforward to read. I recommend with minor revisions.
The strength of the manuscript is using state-of-the-art remote sensing methods like TLS, SfM, and BLS for analysis. The authors also did comprehensive field work and use clear figures, tables to illustrate the data processing and results.
However, there are some aspects that can be improved: (1) the abstract is too long (>500 words). I would suggest rewriting and meet the journal word limit requirement; (2) please double check the terms used in the text and make use they are explained, for example, what is segmentation accuracy (F), what is SLAM (Line 174)? What is ΔRMSE?
Some additional comments are following:
Fig. 1 Please add a background map showing the location of the three example plots.
Fig. 3 please add subfig names: a, b, c,….
Line 239 you mention the volume calculation in the title but not in the text.
Author Response
Please see the word file.
Author Response File: Author Response.docx
Reviewer 3 Report
Dear Authors,
Thanks for this paper.
The paper analyses applicability of photogrammetry and portable laser scanning devices to derive forest stand characteristics at three various (coniferous, broadleaf, mixed) testing plots (30x30m) in subtropical conditions of the university arboretum. Although results are limited to these three testing plots, these provide an overview of comparative measurements at the plots of diverse complexity - simple, medium and complex (number of trees and understory).
Pleae find here few suggestions from the review:
L155 Theare are also thiner woody plants in the plots than indicated in the range provided in the table - please explain and justify applied thresholds.
L191, 192 Please align used terms "large circle", "big circle".
L276 Please provide explanation of interpretation of F in individual tree segmentation, please explain errors in individual trees detection.
L293 Please check texts in figure (Croadleaf), please provide definition of crown boundary in your research (figures) (as it does not correspond to real extent of crows).
L307, 327 Please check texts in figure (extrected)
L444 Please check text description of figure (a1-a3 SfM acquisition route) and camera directions in c-2 (visibility of trees in small circle).
L495-513 Please provide also conclusion on the effectivity of individual approaches - ratio of price of used devices, time consumtion for measurement and processing.
Author Response
Please see the word file.
Author Response File: Author Response.docx
Reviewer 4 Report
This paper tackles an interesting problem, and could be considered a good contribution to the field. My more detailed comments on this study are:
- Introduction could be improved by adding more related studies.
- In the introduction and abstract, one of the limitations of TLS is mentioned to be non-portability. However, there are studies that use portable low-cost TLS systems for evaluating forest structures. What is the advantage of using SFM and BLS instead of those portable low-cost TLS systems?
- "In this study, BLS and SfM point clouds were coarsely registered by rotational translation matrices based on TLS point cloud data. Point cloud rotation, translation, and 208 zoom factors were calculated by randomly selected 20 control points in TLS, BLS, and SfM 209 point clouds"
you need to discuss the registration process in more detail. Registration using these low-density point clouds is one of the more challenging steps of processing these data. How did you overcome the registration bias/error? Discuss more how the rotation translation metrics were derived form TLS.
What do those "randomly selected control points" say about the registration process? Doesn't that add more complexity to the registration. What is the most efficient way to select those control points and how do they affect the registration accuracy?
- Was DBSCAN able to accurately detect the stems without introducing much noise? Based on the point clouds shown in the paper they are relatively complex in terms of point density and structure, so I would assume DBSCAN alone has a very hard time detecting the stems? What other things were done to make this step more accurate?
- How did you address the TLS oversampling bias? Did you use any specific downsampling approach? And why did you choose that method?
- You need to include more related studies in the discussion section and compare their findings with yours. How does your study improve the results/make the process more efficient? What are some steps you can take in the future to address this issue?
- What are the limitations of your method and can it be generalized to other forest environments with more structural complexity? If so, what additional processing tools/steps would be needed?
Author Response
Please see the word file.
Author Response File: Author Response.docx
