Evaluation of Simulation Framework for Detecting the Quality of Forest Tree Stems
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsThe study presents a practical simulation framework concept with valuable potential applications. However, there are significant shortcomings in the arguments regarding simulation fidelity, the explanation of evaluation methods, in-depth result analysis, and the manuscript's presentation standards. The formatting issues are severe. The specific opinions are as follows.
（1）The manuscript’s formatting is severely problematic and needs to be thoroughly revised. All tables must be strictly converted to the standard three-line table format. No formulas should be presented as screenshots. All images must follow a consistent style (size, font, resolution) to ensure clarity and professionalism. Additionally, any orientation errors (such as upside-down images) should be corrected.
（2）The authors should provide a clearer explanation of the innovation of the TCF framework in the introduction and discussion sections. While it is important to acknowledge that the approach is largely based on the integration of existing tools, the paper should emphasize its innovative contributions in terms of system integration, specific problem applications, and customized evaluation methods (EE module). This would help reviewers and readers better understand the positioning and value of the study.
（3） Since the framework is entirely based on simulation data, the degree to which the simulated results align with the real world is crucial. Currently, there is a lack of direct comparison between simulated point clouds (including noise, occlusions, density distribution, etc.) and real harvester LiDAR data. The authors must include verification work in this area, at a minimum providing a detailed comparison plan or a more in-depth discussion of the similarities and differences, to support the validity of the simulation framework.
（4）The weighted evaluation and the EM matrix concept in the EE module are innovative, but the mathematical expressions (Equations 1-8, Table 3 E/A/N) are not reader-friendly. It is strongly recommended to add a concrete, step-by-step computational example and to explain the weight settings and how different error classifications affect the final score in more intuitive language. This will make it easier for readers to understand and replicate.
（5）The description of the SIC module is too vague. The authors need to provide specific parameter values used in key steps (such as ground removal, clustering, and tree validation), the rationale for parameter selection, and any potential sensitivity analysis of these parameters. The absence of these details will affect readers' understanding of the algorithm's performance and the reproducibility of the study.
（6）Given the potential computational bottlenecks in simulation (especially for high-resolution LiDAR simulation), the authors are encouraged to provide information about the approximate computational resource requirements for different simulation settings. Additionally, they should discuss how this framework would perform in various research or development environments (e.g., resource-constrained laboratories vs. large enterprises) and its practical applicability in those contexts.
Author Response
Please check the uploaded response document.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Comments and Suggestions for AuthorsThis study proposed a simulation framework for evaluating the detection of tree stem quality, specifically focusing on curved and defective tree stems. The selected topic is quite an important issue in forestry areas. It helps to improve the efficiency and accuracy of the harvesting systems. However, the paper has several critical points that should be improved. My comments are given as follows;
Only one sentence is written in the abstract about the results. The abstract should reflect the general framework but also provide clear information about the findings.
The text is too long and difficult to follow due to its length. On the other hand, in the 31-page manuscript, the introduction contains very limited information on literature and related-works. In addition to the length of the entire text, the sentences also need to be simplified for clarity.
The Introduction section should clearly define the gaps to introduce the novelty of this study.
Figure 2 is too simple. The proposed approach should be illustrated with a more striking visual.
Even the manuscript mentions that a synthetic forest in a framework can be generated, but there is no sufficient information related to how modeling was realized not provided.
Sufficient information related to point cloud generation by LiDAR should be provided.
Because some of them are partly overlapping, the research questions should be re-arranged.
If heading 2 is Materials and Methods, does your method end at line 194? Headings 3 and 4 should be subheadings of heading 2. Also, there is no need to have that many subheadings. This study can be explained more briefly and concisely than its current form.
Figure 3 is untidy and careless. Figure 9 looks like it was taken directly from another source. Its resolution is very poor. Also, how logical is it to minimize this image after preparing a 31-page article?
Most of the equations were not written by you. You copied them in image format. Equation numbers are also disorganized.
Why the quality heat maps produced by the applied approach belong specifically to classes 5 and 6? Please clarify.
The language of the manuscript should be revised from top to bottom. Additionally, the numbering of the chapters should be re-checked. Some of the chapters have similar numbers, and some of the sub-chapters are not numbered.
There are 35 sources in the references. However, when we exclude internet sources, there are a limited number of references for the 31-page article. The reference section should be strengthened with current studies covering recent times.
Author Response
Please check the uploaded response document.
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsThe following is review of the article titled “Evaluation of simulation framework for detecting the quality 2 of forest tree stems” by Sagar et al. for publication in Forests.
The paper describes a framework to develop simulated Forests which can be used to test or train detection capabilities of LiDAR thus eliminating or at least complementing the use of real-world data. The targeted application is to facilitate integration of LiDAR with harvesting machines to precisely determine the cut extent. I appreciate the authors’ effort in bringing out the study. However, I have the following questions regarding the method and its use.
- While the method is named Tree classification framework, it is mainly doing Tree simulation rather than classification. I suggest changing the method name to a “Forest simulation framework” or so to be appropriate.
- The work is mainly developing a protocol for Forest simulation to be trained with LiDAR image classification. Please illustrate other usable benefits of this method to the broader subject of “Forests” with examples to fit this journal (refer lines 96-98).
- It is described as an open-source tool in the key words, but there is no mention of it later in the text. If making the code open, please add a link to the repository.
- The discussion of optimization methods in lines 80-91seems to have no relation to the paper later. Please remove this.
- Since the suggested application is to be used with harvester machines to precisely determine the cut length, please describe how the method can identify the dimensions of the trees than qualitatively identifying a tree and whether it is curved or straight (in reference in Research question 3).
- In line 131, the method is testing the effect of resolution only, not noise. Occlusion was discussed separately.
- Please clearly define what is the “Ground truth forest” described in the paper. Is it a real forest? If so please give details such as location, area extent, tree density and major tree types.
- Please describe how the rolling terrain is simulated. Is there a measure of surface roughness to it?
- In lines 485-487, please use a different notion other than ‘N’ for no corresponding match category. This confuses it with tree shape classification (S,C,N).
- In lines 634-636, it is described that the LiDAR scan only occurs along one direction(y-axis). Does taking another scan in x-direction improve the results?
- I prefer seeing the RQs to be uniform in section 1 and section 3. But it is up to the authors.
Author Response
Please check the uploaded response document.
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
Comments and Suggestions for AuthorsI thank the authors for the revisions they have made. The authors have responded to almost all of my comments. In general, the revisions made have been found satisfactory by myself. However, despite my previous suggestion in the first round to shorten certain sections, those parts have been extended further but, this can be considered acceptable since it contributes to the overall coherence of the manuscript.
Although Figure 2 has been revised, it is not significantly improved compared to the previous version; its visual quality remains quite poor. On a positive note, the Appendix now includes sufficient and meaningful information regarding point cloud generation by LiDAR.
The code presented in Figure 9 should be presented reformatted as either a flowchart or pseudocode. Alternatively, this code might be placed in the Appendix. The text in Table 3 should be aligned properly (left or right justified). Equations and their numbering should be checked, and corresponding corrections should also be made in the manuscript.
Author Response
Kindly refer to the attached response.
Br,
Anwar Sagar & coauthors
Author Response File: Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for AuthorsI feel the authors have sufficiently addressed my comments. I agree to publish the paper in its current form.
Author Response
Kindly refer to the attached response.
Br,
Anwar Sagar & coauthors
Author Response File: Author Response.pdf
