Next Article in Journal
Detecting the Distribution of Callery Pear (Pyrus calleryana) in an Urban U.S. Landscape Using High Spatial Resolution Satellite Imagery and Machine Learning
Previous Article in Journal
Analysis, Simulation, and Scanning Geometry Calibration of Palmer Scanning Units for Airborne Hyperspectral Light Detection and Ranging
 
 
Article
Peer-Review Record

An Improved Point Cloud Filtering Algorithm Applies in Complex Urban Environments

Remote Sens. 2025, 17(8), 1452; https://doi.org/10.3390/rs17081452
by Guangyu Liang 1, Ximin Cui 1,*, Debao Yuan 1,2, Liuya Zhang 1 and Renxu Yang 1
Reviewer 1: Anonymous
Reviewer 2: Anonymous
Reviewer 3: Anonymous
Remote Sens. 2025, 17(8), 1452; https://doi.org/10.3390/rs17081452
Submission received: 23 February 2025 / Revised: 1 April 2025 / Accepted: 16 April 2025 / Published: 18 April 2025
(This article belongs to the Section Urban Remote Sensing)

Round 1

Reviewer 1 Report (Previous Reviewer 4)

Comments and Suggestions for Authors

Compared with the first submitted version, this manuscript has undergone substantial revisions and basically addresses my concerns. It can be accepted for publication.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report (Previous Reviewer 1)

Comments and Suggestions for Authors

(1) Although the paper uses the ISPRS public dataset and data from the NRW region of Germany, the dataset types and coverage scenarios may still not be comprehensive enough. Faced with diverse urban environments, such as mountainous cities, coastal cities and other areas under the influence of special landforms and climate, the algorithm performance has not been fully verified, which limits the evaluation of the universality of the algorithm.

(2) Only the accuracy indicators of different algorithms are compared, and the time complexity of the algorithm is not deeply analyzed. When processing large-scale point cloud data, time cost is a key factor in measuring the practicality of the algorithm. Without this part of the analysis, it is difficult to fully evaluate the efficiency of the algorithm in practical applications.

(3) The paper mentions setting parameters in the NRW experimental area (such as RI=1, GR=1.5m, h=0.3m, etc.), but does not fully explain the optimality of these parameters. Although a sensitivity analysis was conducted (such as the smallest error when GR=1.5m), it did not explain why a specific threshold was selected (such as angle 6°, distance 0.8m), and there was a lack of verification of the generalization ability of parameters in different scenarios.

Comments on the Quality of English Language

The English could be improved to more clearly express the research.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report (New Reviewer)

Comments and Suggestions for Authors

I am not seeing the contribution of this paper:
1.    There are many techniques to identify and segregate the ground plane. Perhaps the authors’ method is extremely fast, but I do not think so because of the TIN, and the information is not reported
2.    The technique requires a fairly flat ground plane, which is a major drawback

Some smaller comments are provided below
For datasets 1-12 please report in a table the point density and total number of points in each
Fig. 2 there is a typo “Extracte”
Fig. 2 the first 3 steps are not sufficiently labeled to be understood without extensive interrogation of the text.
Fig. 2 has 2 different font sizes; please correct that
Fig. 2 the placement of the text for step 3 makes the image very hard to follow. Please reorganize things.
Fig. 2 and accompanying text mentions “random sampling”, but the authors have done spatial sampling or maximal spatial sampling. This is an important error that needs to be corrected. Doing this fundamentally changes the characteristics of the data. This makes me very concerned, especially as there is no data on the point cloud size “before and after”
Table 2 looks terrible. Please reformat it for consistency
Remove Table 3. The techniques are so old that a comparison is non-sensical
For Table 4 to be useful you must report the computing time (either total or average is fine at the bottom of the table as an extra row)
Fig. 9 has too much on it. It cannot be understood as currently configured. Try using a color progression for the column chart data instead of jumping around as no patterns can be established with the current color assignments
The data in Fig. 10 and Fig. 11 and Fig. 13 are not consecutive. So you cannot use this graph time
Fig. 12 has typos “angle” not “angel”

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 4 Report (New Reviewer)

Comments and Suggestions for Authors

I see that the article was corrected after review. My  recommendations:

  1.  (https://www.bezreg- 104koeln.nrw.de) -  add the Intenet site to reference list;
  2. Figure 4. The experimental area was preprocessed and projected in a plane according to height. - explain all side data. What shows the colours?
  3. Table 2. - explain the matrics of calculation. 
  4. I do not find the information about the samples (Table 3).
  5. The list of references is suitable.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report (Previous Reviewer 1)

Comments and Suggestions for Authors

The author has responded most of my concerns, and I recommend that the paper can be accepted after carefully examining the writing of the paper and polishing the language.

Comments on the Quality of English Language

The English could be improved to more clearly express the research.

This manuscript is a resubmission of an earlier submission. The following is a list of the peer review reports and author responses from that submission.


Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

1. The experimental data is only sourced from the state of North Rhine Westphalia in Germany, and the geographical limitations of the data may result in insufficient validation of the algorithm's applicability in other regions or different terrain features.

The parameter settings in the algorithm (such as an angle threshold of 6 ° and a distance threshold of 0.8m) are determined through experimental analysis, but in practical applications, different data characteristics may require different parameter combinations, and dynamic parameter adjustment can be considered.

3. Insufficient comprehensive comparison with other latest related research methods

Comments on the Quality of English Language

The English could be improved to more clearly express the research.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The paper presents a novel approach to point cloud filtering. The authors use cloth-simulation-based algorithms and triangulated irregular network models in urban scenarios for ground extraction. Below are my comments for the authors:

 

# Strengths

- Novel point filtering approach

 

# Weaknesses

- A bit weak evaluation

 

Major comments:

- I would recommend that the authors investigate why Sites 1 and 11 have a similar number of Type 1 errors of CSF to Type 2 errors.

 

Minor comments:

- In Figure 14, it is not clear how (a) and (b) are comparable. While one image shows point classifications, the other shows the triangulation of one class. It would make sense to show comparable content in the images.

- I miss the standard deviation of the results in Table 4.

- TIN is used before defined in the Abstract

- CAP is introduced before it is explained

- page 3, section 2.1 ... there is no need to describe LAS format in such detail. It is a common format and does not need additional presentation

- There is space missing in "Table4" label

- Titles of subsections are hanging on previous pages (2.5.2., 3., 3.1, 3.2, 4.2, )

 

Taking into regard the above comments, I recommend a minor review of the paper.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

Please see the attachment.

Comments for author File: Comments.pdf

Comments on the Quality of English Language

Improvement in English proficiency is advised.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 4 Report

Comments and Suggestions for Authors

The article presents a point cloud filtering algorithm based on cloth simulation and the TIN model (CAP). Experimental results demonstrate that the proposed algorithm performs well in terms of accuracy and adaptability for point cloud filtering tasks in urban environments, showing significant potential for practical applications. Overall, the manuscript is well-structured, and the experiments are thorough. With minor revisions, it can be accepted. The following suggestions are provided for consideration during revisions:

 

1. It is recommended to separate the literature review from the introduction. In research papers, the introduction and the literature review are typically independent. The point cloud processing algorithms could be categorized into traditional methods and deep learning-based approaches. 

 

2. Since the foundational research data is based on publicly available datasets, it would be beneficial to also provide links to the code for the proposed algorithm and the comparison algorithms. This would facilitate better replication and comparison of the algorithm's performance.

 

3. The conclusion focuses on the three experimental results of CAP and the corresponding advantages; however, the drawbacks and limitations of CAP (such as potential issues with robustness in filtering different types of targets) are not addressed. It would be helpful to include these in the conclusion for a more balanced discussion.

 

4. Since the experiments use publicly available data, section 2.1 might be overly detailed. There may be no need for an extensive introduction to the public dataset.

 

5. Sections 2.4 and 2.5 contain a common issue, similar to the descriptions in Figures 6 and 7. The authors only qualitatively describe how points are selected and edges are connected, without providing a specific procedure. This is not very user-friendly for readers attempting to replicate the method. It is recommended to revise these sections by providing specific algorithmic steps, perhaps illustrated with examples and diagrams of particular points and edges.

 

6. Some formatting issues exist in the manuscript, such as inconsistent font usage in section 4.3.

 

7. In the experimental section, the authors provide a detailed analysis of various parameters for the proposed algorithm's modules, but I did not notice any comparison with other algorithms. Would this be appropriate? As it stands, the paper resembles more of an experimental report than a comprehensive research study.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

Please see the attached document

Comments for author File: Comments.pdf

Comments on the Quality of English Language

Please see the attached document

Back to TopTop