Automatic Center Detection of Tropical Cyclone Using Image Processing Based on the Operational Radar Network

Round 1

Reviewer 1 Report

Automatic center detection of tropical cyclone using image processing based on the operational radar network

 

Review report

 

The paper is based on determining the center of tropical cyclones using image processing techniques. Here the method utilizes the concept of locating the maximum circulation point of the TC center using a motion vector field derived from changes between radar images. The algorithm is the Automatic Center detection of Tropical cyclone using Image processing based on Operational radar Network, abbreviated as ACTION. 

 

General comments

The paper is written in a way such that it is difficult to follow the flow of the algorithm. In addition to that the details of the radar variable and the data that is used is not very clear. Lot of data has been used in the analysis to determine the TC center which is a big advantage of this paper. The accuracy of the detected TC is also very high which is very encouraging. The detail as to whether the radar is ground based (I am assuming that it is ground-based network) or airborne is not clear. Therefore, the paper needs to be modified a lot before it can be accepted. The introduction is too short and needs to include more details and background study.

 

Detailed comments

Abstract: ACTION easily calculates using Open Source Computer Vision, performs in real time, and can be directly applied to rapidly generated weather radar images; hence, it is currently being utilized by the KMA.

The reviewer is not sure why this sentence is so strangely written and what is the reason for mentioning Open Source Computer Vision.

 

The authors should try to use proper scientific terminology. The use of words like dangerous, clue should be avoided. 

 

Line 64: The derived vector field is performed through computer image processing in which the forecaster selects the center point of rotation of the TC as the center of the TC in the radar echo image, minimizing the subjectivity of TC analysis through the radar image among forecasters.

The sentence is very long and difficult to follow and strangely written.

 

Page 3. “cv2.calcOpticalFlowFarneback” details can be included as an appendix. It is not usual practice to include these details in the main sections.

 

Line 138-139: The radar variables have been introduced here but there is hardly any discussion on the use of the variables in this study. Even the discussion on figure 3 does not have any information about reflectivity or differential reflectivity. 

 

Line 149: In the first step, basic motion vectors are extracted from continuous radar images through the optical flow technique. There is no discussion about the variable on which the image is based on.

 

Line 150-170: There is an extensive discussion on the resolution required to determine the u and v components. However, there is no discussion on the variable that was used for this study.

 

Whether the radar is ground based or airborne? There is no discussion on that either. 

 

Line 191-192: To minimize this problem in the algorithm, some vector blank areas of the radar echo proximity region are linearly interpolated using the neighboring vector data. 

Does the accuracy of the u, v component get affected by this interpolation. What is the quantitative impact?

 

Line 216-217: To create such a dense vector field, the resolution was increased from 40 to 20 km, and the vector empty area was interpolated using the radial basis function (RBF) to generate the dense motion vector field for the entire analysis field. It has been shown that 40 km window size is appropriate and lower window size leads to errors. Here again the authors want to use a dense vector field and smaller window size. This is confusing.

 Authors should use a flowchart to explain  the algorithm. As it is, the algorithm blocks seem all distributed due to which the reader might not get a full understanding of the process.

 

The authors should be careful about the use of capital letters. They have used capital letters at random places in a sentence.

Labels of figure 4 should be increased in font size.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

This paper presents the algorithm allowing to detect motion of the TC’s center from a weather radar imagery. The method demonstrates a high detection rate and a good spatial proximity of the derived TC’s track to the best track recorded by the Korea Meteorological Administration (KMA). The method was patented and, according to authors, is utilized by the KMA.

However, in my opinion, the manuscript doesn’t fit the academic standards and is not appropriate for publication in Atmosphere. Generally, it describes results of the applied engineering work, which are already known (patented). Moreover, extensive editing of English language and style is required. There is also an obvious error - Eq.(1) is formally inconsistent with Eq.(2). This error doesn’t affect the algorithm and its applications. Yet, it is disturbing to see this obvious inaccuracy in the manuscript.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

I'd recommend an additional editing of English language and style