A New Method for Positional Accuracy Analysis in Georeferenced Satellite Images without Independent Ground Control Points

Round 1

Reviewer 1 Report

1. The biggest improvement of the paper is that the control points are not obtained by field survey？It is mentioned in the paper that the plane coordinates of the control points are provided by high-precision DOM, and the elevation can be obtained by DEM Interpolation. Then the question to be answered in this paper is that the accuracy of control points provided by this method can reach the same accuracy as field survey. And What is the innovation of this paper?

 

2.“Most importantly, our method is more efficient computationally. Our experimental results confirm that the method is equally effective and less expensive than other available methods to analyze the positional accuracy of satellite images.”

If the author has such a conclusion, please give a quantitative description in the abstract.

 

3.”Specifically, it recommends that the accuracy of the control data should be at least three times the accuracy of the analyzed data. “

What is the theoretical basis for this three times the accuracy of the analyzed data？ Please describe it.

 

4. “As was previously stated, the xy-coordinate pair for each pixel is located inside the pixel limits. “

This sentence is not clear. Please explain it.

 

5.Do not understand the meaning of each number in Table 2. There is no unit. What's the point of doing 30 groups of experiments? What is the basis for the choice of these 30 groups of experiments?

 

6.”To better understand these values, we must stress an important difference with regards to the sample size in both approaches: in the traditional approach, a sample size of 19 IGCP points was used, while a greater, more representative sample size of 39 ICP points was used to measure the positional accuracy.”

What is the difference between the traditional method and the paper’s method? Only automatic matching and uniform distribution of points by PDS? Is it possible to use grid in the image to achieve approximately uniform distribution? In theory, four GCPs are enough for one image orientation.

 

7. Please describe the geometric model and orientation method used in this paper.

 

8. There are too many text descriptions in this paper, which should be represented by tables, such as input image data, ground control and inspection data and other parameter information.

Author Response

The authors gratefully acknowledge the Associate Editor and the three Anonymous Reviewers for their detailed and highly constructive criticisms, which greatly helped us to improve the quality and presentation of our manuscript. In the following, we provide detailed, item-by-item, point-by-point responses to all the very interesting issues raised by the Anonymous Reviewers. Please note that, in order to simplify the review of our manuscript, we have highlighted the main modifications introduced in the revised manuscript in blue color to help the Editor and the Anonymous Reviewers in finding the changes made with regards to the previous version. We are indebted to them for their careful assessment and outstanding suggestions for improving our manuscript, which have been really helpful in order to enhance its presentation and technical quality.

Author Response File: Author Response.docx

Reviewer 2 Report

The authors have developed a new method  for Positional Accuracy Analysis in  Georeferenced Satellite Images without Independent 3 Ground Control Points.

The manuscript faithfully follows the structure of the scientific method and each of its sections has been carefully written.

The introduction has been divided into 3 sections: introduction, related works and main contribution. This way it is possible to clearly differentiate the state of the art of the research and the main new contribution that this new study brings. However, I believe that the previous bibliography concerning this field of knowledge has not been studied in depth.

In general, the whole manuscript denotes a great understanding of that field of knowledge. However, in my opinion, the proposed development is interesting but does not represent a great advance over previous methods.

Nevertheless, I believe that the manuscript is suitable for publication in Remote Sensing, once some minor revisions have been solved.

 

Minor revisions:

• Increase the bibliography referring to the state of the art in this field of knowledge
• Figure 1 should show the position coordinates according to Remote Sensing journal standards

Author Response

The authors gratefully acknowledge the Associate Editor and the three Anonymous Reviewers for their detailed and highly constructive criticisms, which greatly helped us to improve the quality and presentation of our manuscript. In the following, we provide detailed, item-by-item, point-by-point responses to all the very interesting issues raised by the Anonymous Reviewers. Please note that, in order to simplify the review of our manuscript, we have highlighted the main modifications introduced in the revised manuscript in blue color to help the Editor and the Anonymous Reviewers in finding the changes made with regards to the previous version. We are indebted to them for their careful assessment and outstanding suggestions for improving our manuscript, which have been really helpful in order to enhance its presentation and technical quality.

Author Response File: Author Response.docx

Reviewer 3 Report

General Comment

 

The submitted paper introduces a semi-automatic technique to evaluate the positional accuracy of each pixel in a georeferenced satellite image without the need for identifying physically surveyed ground control points in the study area on the satellite image.

In general, the submitted manuscript is well organized. Comments about some statements in the current manuscript are given as below.

 

Comment 1

Please provide an appropriate citation for the statement in lines 108-110.

Comment 2

Please clarify the spatial resolution of the selected panchromatic band in the satellite image in the statement in line131.

Comment 3

Regarding the information provided in the subsection 4.2. (Positional Control Data), the ortho-image which is used as the control area in this study has been generated using topographic information extracted from LiDAR data with mesh size 5 by 5 meter and the positional accuracy with RMSE of 2 meters. According to the fact that this positional accuracy is lower than the positional accuracy of each image pixel in the aerial images with the planimetric accuracy of ±50 cm it will negatively affect the ultimate positional accuracy of the derived ortho-image. However, it seems that the authors are referring to the original planimetric accuracy of the aerial imagery throughout the manuscript. The authors are required to clarify this issue.

Moreover, the authors should discriminate and report the error budget and/or error statistics on each of three directions (X, Y, Z) over sampled ICPs and IGCPs.  

Comment 4

Please provide what PDS stands for in line 204, where you introduce this acronym for the first time.

Comment 5

In lines 207-208, “To comply with the ASPRSb requirement, …”, what does the authors mean by saying that the CA resolution from which we will extract must be at least 3 times more accurate? If the authors mean the positional accuracy of the image pixels, they should rewrite the statement above more clearly because the accuracy in resolution does not make any sense in this context.

Regarding lines 209-214, the authors should notice that having aerial imagery with finer spatial resolution does not necessarily mean higher spatial accuracy. The spatial accuracy of each image pixel is significantly related to the accuracy of the method employed for geolocating the aerial image.

Comment 6

Please provide an appropriate citation for the statement in lines 214-215.

Comment 7

In the proposed methodology, it is not clear enough to the potential readers why the image positions of the sampled points for providing  ICP_CA and corresponding ICP_SA, have been rounded (or reduced) to the coordinate of the center of the pixel covering the sampled points in CA and SA by PDS algorithm while extracting those xy-image coordinates with subpixel accuracy can result in more accurate analysis on determining the positional accuracy of the satellite image. Authors are recommended to clarify what is the reason behind ignoring the determination of ICP_CA and ICP_SA with subpixel  accuracy.

 

Comments for author File: Comments.docx

Author Response

The authors gratefully acknowledge the Associate Editor and the three Anonymous Reviewers for their detailed and highly constructive criticisms, which greatly helped us to improve the quality and presentation of our manuscript. In the following, we provide detailed, item-by-item, point-by-point responses to all the very interesting issues raised by the Anonymous Reviewers. Please note that, in order to simplify the review of our manuscript, we have highlighted the main modifications introduced in the revised manuscript in blue color to help the Editor and the Anonymous Reviewers in finding the changes made with regards to the previous version. We are indebted to them for their careful assessment and outstanding suggestions for improving our manuscript, which have been really helpful in order to enhance its presentation and technical quality.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Compared with the previous version has a significant improvement, hope to be published as soon as possible. 

Reviewer 3 Report

Thank you for your response.