Extracting Individual Tree Positions in Closed-Canopy Stands Using a Multi-Source Local Maxima Method

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Estimating the position of individual tree in forest stands has always been an important part of forest inventory, particularly for assessing stand structure in closed canopy environments. The use of Unmanned Aerial Vehicle (UAV) enhances the precision of this task. While the concept and methodology presented in the research are not novel, certain methodological approaches are valuable. The study enhances our understanding of the Revised Local Maxima (RLM) method and introduces a Multi-Source Local Maxima (MSLM) method based on UAV-visible light data. Field surveys, UAV data, and the MSLM method are proposed to enhance the accuracy of individual tree location across various forest stand types in the Jinpen Mountain Forest Farm. The overall structure and style of the paper leave a positive impression. Given the field of study, methodology, and results obtained, this research aligns well with the scope of the journal "Forests."

I have only several minor concerns:

40-42 / 92-96 /129-130 I recommend adding some references here.

158 Generated maps (data)?

159 Study methods

169 [30] I’m not sure that this reference is relevant in this context.

187 Flowchart should be sharper

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

I have thoroughly reviewed the study titled 'Study on Extracting Individual Tree Positions in Closed-Canopy Stands Using a Multi-Source Local Maxima Method' and have identified several fundamental issues that require attention. While the subject matter and research focus are intriguing, the manuscript's writing style lacks the necessary scientific rigor and clarity. In terms of novelty, the application of the MSLM method presents a valuable enhancement; however, its justification and positioning within the existing literature need to be strengthened. Below, I provide a detailed critique addressing both methodological and presentation-related concerns.

• I think the phrase "Study on" in the title is unnecessary. A more concise and professional title suggestion to the authors would be "Extracting Individual Tree Positions in Closed-Canopy Stands Using a Multi-Source Local Maxima Method."
• The Introduction lacks a clear justification for the necessity of the proposed MSLM method beyond its improvement over the RLM method. To strengthen the manuscript, the authors should explicitly identify the gaps in the existing literature that MSLM addresses, such as the key limitations of current tree detection approaches. Additionally, a more in-depth discussion is needed to explain why integrating CHM and DOM is advantageous compared to other data fusion techniques.
• The final paragraph of the Introduction should be structured to clearly outline the rationale for the study, the research hypothesis, the study question, the specific gaps it aims to fill in the literature, and its key objectives. However, the current version does not sufficiently address these critical elements. To enhance clarity and scientific rigor, the authors should thoroughly revise this section, ensuring it effectively communicates the study's significance and contributions to the field (Lines 82-89).
• The study design includes 1,215 trees across nine plots in a single location. While this sample size appears adequate for the scope of the study, its generalizability could be significantly improved by incorporating more diverse forest compositions (mixed stand vs. pure stand), varying canopy densities (sparse vs. dense forests), and different terrain conditions (flat vs. mountainous landscapes).  Since the study has already been conducted, a redesign is not feasible at this stage; however, I strongly recommend that the authors acknowledge this limitation in the discussion and consider incorporating a more diverse dataset in future research to enhance the robustness and applicability of their finding (Lines 101-110).
• The reported accuracy metrics do not include confidence intervals or statistical significance testing. Authors must include confidence intervals (e.g., 95% CI) for accuracy metrics.
• Several figures (Figure 2 and Figure 5) lack clear legends, scale bars, and units. Authors should add scale bars and unit labels to all figures and should use consistent color schemes across figures to improve readability. It is tough to read some of the figures.
• Some figures are not referenced before appearing in the text; ensure that all figures are cited before appearing in the manuscript (Figures 1, 2, and 4).

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

The manuscript deals with the detection of individual trees positions in a dense canopy using drone RGB images. The paper is generally well prepared. I have only some notes about the description of methods.

·       As forests are plantations, apparently, they are even-aged. So, can you give the age, mean DBH and height for sample plots? Are trees planted in a regular grid (in this case determining trees position is a trivial task) or stochastically?

·       It should be explained more clearly the difference between DOM and DSM, DSM and DTM. DOM is original photo in RGB colors, isn’t it? Which particular variables are attributed to each pixel in each case?

·       It is not enough clear how did you moved from 2D RGB images to 3D point cloud. As I understood, you did not use LIDAR, only RGB photos. Is it somehow based on the combining of views of the same point from different angles/directions?

·       In 2.3.1. it is not clear why Grey maximums must correspond to crown apices. The highest Grey values correspond to the most dark points, isn’t it? Why must these points be crown apices? It is more reasonable to expect that darkest points are in the shadow.

Other notes are in the attachment.

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have reviewed all the points I suggested in my previous review one by one and made the necessary improvements. I think the study can be accepted in its current form.