Tree Skeletonization with DBSCAN Clustering Using Terrestrial Laser Scanning Data

Round 1

Reviewer 1 Report

An excellent and very well-written and structured paper. The rationale for the research is clear, there is good coverage of the literature, and clear presentation of the method etc. Maybe the Introduction could be shorter and the coverage of the literature separated out into a Background/Literature Review section. This might allow additional clarity in the Introduction about the focus and purpose of the research, leaving the Background section more opportunity to highlight what has been done, the basis for the research in the Background? Just a thought. Also, perhaps a little more context at the start and in the conclusions about the potential applicability of the approach in relation to the need to be able to find non-destructive ways to 'map' the tree canopy structure..... perhaps use in modelling of tree canopies e.g. for reflectance etc... Any possibilities for the use of drone technology carrying Lidar? But otherwise very informative and well explained.

Author Response

Please see the attachment. 

We hope our modification and explanation is satisfactory. Thank you!

Author Response File: Author Response.docx

Reviewer 2 Report

In this study, authors proposed a tree skeletonization method based on density-based spatial clustering of applications with noise (DBSCAN) using TLS data. 

They proposed and tested followings steps:

step 1:  outliers wereremoved using DBSCAN and the point traversal order of each point was recorded. 

step 2: a tree point cloud was divided into several tree slices by contour planes, and several tree segments were obtained by applying DBSCAN to each tree slice.

step 3: tree skeleton points were retrieved from each tree segment after the point inversion transformation. 

step 4: the adjacent relationship between skeleton points and the flow weight of each skeleton point were calculated based on the point traversal order. 

step 5: the skeleton points were classified into stems and different levels of branch points by the flow weights of skeleton points, and the branch hierarchies were identified. 

step 6: the tree skeleton was optimized by the angle consistency.

 

The idea of tis paper is good but a methods are well known.

Density-based spatial clustering of applications with noise (DBSCAN) is a data clustering algorithm proposed by Martin Ester, Hans-Peter Kriegel, Jörg Sander and Xiaowei Xu in 1996. The algorithm was awarded the many award ant its still testsed and developed.

The pseudocode of the algorithm can be even found in Wikipedia.

Although the results seem satisfactory, there is no reference/comparision to another method of tree skeletonization.

On page 13 line 311 authors wrote: "This means that the method in this study needs to be further improved." In Discossion section authors wrote about futher work with proposed method. 

So maybe they should try to do their plan first to definitely improve the paper.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

 

I think that after the answers given, as well as my and other reviewers' comments, the paper can be accepted.

However, I believe that the paper still lacks a clearly demonstrated existing approach and a new/modified method. Authors can do this in the form of a block diagram or showing a new part of the code.

Authors’ comment:

Unfortunately, we did not find the source code of these references, nor can we make a graphical comparison. From the perspective of tree height, complexity and tree structure composition involved in these references, the proposed method has some advantages.

Answer:

DBSCAN - Wikipedia

https://en.wikipedia.org/wiki/DBSCAN

(please find the appendix with print screen from Wikipedia) 

Comments for author File: Comments.docx

Author Response

Please see the attachment.

Author Response File: Author Response.docx