KNN Based Denoising Algorithm for Photon-Counting LiDAR: Numerical Simulation and Parameter Optimization Design

Round 1

Reviewer 1 Report

This paper presents a KNN based denoising algorithm for photon-counting LiDAR, where the important parameters are adjusted according to the background photon count rate. The introduction states the purpose of the paper, and the relation between the paper and the previous works is clearly explained. The improvement of the presented algorithm compared with other classical algorithm is illustrated via experiments.

 

Overall the paper is interesting and clearly written. The following are my specific comments:

 

(1)  In Section 2, how is the moving speed of the light spot on the ground taken into considered in the tuning of the value of k and the threshold selection? It is better to provide some explanations.

(2)  Figure 3 and Figure 4 are not cited in the text.

(3)  The theoretical analysis in Section 2 is not satisfactory. The efficiency of the presented algorithm is not verified in theoretically.

(4)  I wonder whether the optimal range of k = 2-6 is determined for certain scenarios? Why is the k value defined as 50 through empirical judgment? The authors are suggested to provide some suggestions for the application of the optimal range.

(5)  How is the scale factor q that is changed with the background in equation (9) determined? A practical approach for the calculation of q should be given.

(6)  The function between the threshold and the background photon count rate is obtained based on the simulated noise signal. Try to provide some remarks about the effect of the model error to the performance of the algorithm.

(7)  In Section 4, in addition to figure 13 and figure 14, some statistical values should be provided to illustrate the high performance of the presented algorithm.

Author Response

Dear reviewers,

We would like to thank reviewers for the positive and constructive comments concerning our submitted manuscript. Those comments are all valuable and helpful for revising and improving our paper. We have studied the comments carefully. According to the comments, we have made corrections which we hope meet with your approval. In the revised manuscript, all corrections are marked (in red font). The responses are as follows (in blue font).

I hope this revision can meet your expectations.

Author Response File: Author Response.pdf

Reviewer 2 Report

 

The paper presents a point cloud data denoising algorithm. The echos data was generated by simulation based on probability distribution functions.

 

Here are the comments:

First, it is not clear why the approach considers 2d data by generating 2d profile point cloud data rather than using the more informative 3d point data?

 

Equation (2)  is referring to reference [21] but it is not explicit to make a direct link. The authors should clarify the context.

 

Also, the steps from equation (2) to equation (3) are not clear?

 

Is there a difference between hitting a photon and detecting k photons in line (174) ?

 

In figure 7, i and j are inversed or k=i-j should be k=j-i in line 276 ?

 

 

In table 1, there are NaN values with the KNN approach. What is the reason? 

 

 

My concern is about using simulated signal which is also converted to a 2d profile. The question that may arise is why in this work the authors did not consider real data as a homemade UAV was used.

 

I guess that it is more interesting to use a well-known ground truth scene and then remove the noise from the signal by evaluating the reconstructed 3d signal? 

Author Response

Dear reviewers,

We would like to thank reviewers for the positive and constructive comments concerning our submitted manuscript. Those comments are all valuable and helpful for revising and improving our paper. We have studied the comments carefully. According to the comments, we have made corrections which we hope meet with your approval. In the revised manuscript, all corrections are marked (in red font). The responses are as follows (in blue font).

I hope this revision can meet your expectations.

Author Response File: Author Response.pdf

Reviewer 3 Report

The paper is well defined and well organized. There is just little english correction is required. Congratulation.

Author Response

Dear reviewers,

Thank you very much for encouraging comment. We have carefully checked the paper and revised some of the content. Thanks for your comments for improving this article.