Next Article in Journal
Analysis of the Expression Patterns of 13 DREB Family Genes Related to Cone-Setting Genes in Hybrid Larch (Larix kaempferi × Larix olgensis)
Next Article in Special Issue
Sliding Cutting and Cutting Parameters of Concentric Curvilineal Edge Sliding Cutter for Caragana korshinskii (C.K.) Branches
Previous Article in Journal
Water Use Efficiency of Five Tree Species and Its Relationships with Leaf Nutrients in a Subtropical Broad-Leaf Evergreen Forest of Southern China
Previous Article in Special Issue
Remote Monitoring of Amur Tigers in Forest Ecosystems Using Improved YOLOX Algorithm
 
 
Article
Peer-Review Record

A Deformable Shape Model for Automatic and Real-Time Dendrometry

Forests 2023, 14(12), 2299; https://doi.org/10.3390/f14122299
by Lucas A. Wells *,† and Woodam Chung
Reviewer 2:
Reviewer 3: Anonymous
Forests 2023, 14(12), 2299; https://doi.org/10.3390/f14122299
Submission received: 9 October 2023 / Revised: 11 November 2023 / Accepted: 21 November 2023 / Published: 23 November 2023
(This article belongs to the Special Issue New Development of Smart Forestry: Machine and Automation)

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Dear Authors,

I want to thank the authors of this research first of all. Their case study details the steps involved in creating and getting ready a fairly sophisticated algorithm in accordance with forestry requirements. At this point, I'd like to make clear that while this article's explanation of a portion of a patent-registered piece of software is comparable to the authors' earlier works, it differs from them in terms of content. It is undeniable that an already-in-use algorithm will significantly advance the field of forestry research. Neither the article's language nor its content can be faulted. There is only one aspect of the study that can be considered as improvable. This is during the accuracy analysis step, 21 tree trunks were examined. Although the study may be improved in this area, the results are still deemed satisfactory because a remarkable amount of work was put into developing the algorithm. In addition, a few recommendations might be made from the perspective of the reader to make it simpler for those who are interested. A graphical representation of the algorithmic path and an indication of the work's mainstages will improve comprehension at this point. A graphical abstract would make the article appear more appealing at first sight.  

Best regards.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

In this study, an stereo image based algorithm is introduced for measuring tree trunk diameter; it is stated that the algorithm can automatically estimate the diameter of trees up to 20 meters away with a 5.52% margin of error

The methodology introduced in this study exhibits potential for selective implementation across diverse forestry operations. As delineated in the concluding remarks, the technique may not consistently manifest optimal outcomes in natural forest ecosystems characterized by stratified arboreal forest layers or understory vegetative density. An empirical validation of the algorithm under authentic forest conditions, beyond park-like settings, would enhance the robustness and broader applicability of our findings.

However, despite some limitations, this approach is valuable as a simple method that can be integrated into mobile/etc.. applications for forestry operations.

PS: double 'the' in line 541

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

The idea is good. However, I are very suprised that the proposed method is not be compared in experiments. Then, how to present the advantages of this work.  There is no researcher on this area?

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Back to TopTop