Validation of Close-Range Photogrammetry for Architectural and Archaeological Heritage: Analysis of Point Density and 3D Mesh Geometry
Round 1
Reviewer 1 Report
Dear,
the work is well structured even if it does not present an absolutely new study. I believe, however, that it is necessary, from time to time, to pay attention to the comparison between LS and SfM methods to remind people of the pros and cons of the two systems.
However, I would like to read in the conclusions some proposals for the improvement of the result and, above all, a conclusion of the paragraph "conclusions" worthy of a conclusion (pardon for the word-pun). In fact, the speech is suddenly interrupted.
Author Response
Response to Reviewer 1 Comments
Point 1: The work is well structured even if it does not present an absolutely new study. I believe, however, that it is necessary, from time to time, to pay attention to the comparison between LS and SfM methods to remind people of the pros and cons of the two systems.
Response 1: We appreciate the Reviewer’s suggestion. We thank the anonymous Reviewer for her/his time spent suggesting improvements for our paper.
Point 2: However, I would like to read in the conclusions some proposals for the improvement of the result and, above all, a conclusion of the paragraph "conclusions" worthy of a conclusion (pardon for the word-pun). In fact, the speech is suddenly interrupted.
Response 2: We appreciate the Reviewer’s suggestion. Conclusions have been modified and extended.
Reviewer 2 Report
I am afraid I stopped commenting around page 13. I think the paper needs simplified. I see multiple cameras, multiple software, multiple instruments, multiple algorithms, multiple vague terminology with no clear rationale it seems to me. The methodology needs to be presented in a clear, unambiguous manner; what did you do, how did you do it, why did you do it, when did you do it and so on? If the methodology is not clear and robust neither will its outputs be. Please spell it out in a manner that a wayfaring man, though he were a fool cannot err therein so as to speak.
I do not wish to appear unkind or ungracious, but, at times, I felt like I was staring through fog, I am sure there is something on the other side, I am not just sure what at this stage.
The literature review seems to extend throughout the paper, however, it is not enough to simply, and uncritically, quote authors in support of your arguments. What does the literature show, what are the gaps in the literature and how does this paper address them?
On a positive note, having scanned the output data, I think there is more than one paper in this study. Think of rigour, think of significance, think of impact.
I have a marked up pdf that I am happy to share with the authors if they wish and I hope that my comments here and in it, are helpful.
Best wishes.
Author Response
Response to Reviewer 2 Comments
Point 1: I am afraid I stopped commenting around page 13. I think the paper needs simplified. I see multiple cameras, multiple software, multiple instruments, multiple algorithms, multiple vague terminology with no clear rationale it seems to me. The methodology needs to be presented in a clear, unambiguous manner; what did you do, how did you do it, why did you do it, when did you do it and so on? If the methodology is not clear and robust neither will its outputs be. Please spell it out in a manner that a wayfaring man, though he were a fool cannot err therein so as to speak.
I do not wish to appear unkind or ungracious, but, at times, I felt like I was staring through fog, I am sure there is something on the other side, I am not just sure what at this stage.
Response 1: We appreciate the Reviewer’s suggestion. We are grateful for your sincerity and the positive note on the fact that the material of this study could be used for more than one paper. Honestly, we are sure that the work of researchers is increasing the knowledge in their area, and publishing in a high impact journal is always a challenge. For this purpose, it is sometimes required to carry out many tasks so that the research study is completely valid. Regarding the methodology, the authors have structured the paper according to an applied analysis approach.
Point 2: The literature review seems to extend throughout the paper, however, it is not enough to simply, and uncritically, quote authors in support of your arguments. What does the literature show, what are the gaps in the literature and how does this paper address them?
Response 2: We appreciate the Reviewer’s suggestion. Regarding your comment on quotations, we think that it is the best way to justify a research study. It is interesting to present discussion notes in the results and explanatory notes in the methodology to explain work processes without excessively repeating the aspects already published. We understand that quotations could be included in the conclusions to corroborate the research study, but sometimes they are not included, according to the general opinion of the scientific community.
Point 3: The On a positive note, having scanned the output data, I think there is more than one paper in this study. Think of rigour, think of significance, think of impact.
I have a marked up pdf that I am happy to share with the authors if they wish and I hope that my comments here and in it, are helpful.
Response 3: We appreciate the Reviewer’s suggestion. Based on your comments, the results have been modified and the conclusions have been extended. We hope that these modifications are appropriate.
Reviewer 3 Report
This paper compares the CRP with the laser scanners. Due to the significant improvement in digital cameras, computation powers, and price, it is clear that CRP and other image based scanning methods are going to be really popular.
While this paper provides the comparison of this methods, this paper is not ready for submission yet:
1- The paper is 21 pages with out even one equation. The writer needs to provide in details that how each of these methods work. What are the fundamentals of each methods, and what are the general pros and cons of them, before even diving into the main process. The authors used these methods like a black box while these methods are described in details in lots of researches before. I don't think this paper is publishable with out covering the mathematics, and fundamentals of each methods in details.
2- The result of imaged based method can be vastly improve via different lens, even different sensor, and proper calibration. I would like to see the result for this parameters.
3- The images based method has multiple steps, such as feature extraction (sift, surf,...), feature matching, epipolar connection, RANSAK,... each of this steps serve as a hyperparameter that can change the final result significantly, I would like to see the result of these parameters.
4- Figure 2 is not informative at all.
5- The x limits axis in both plots in figure 10 need to be cropped
6- Figure 11 is not clear at all.
7- The laser scanner also has its own challenges to improve the result, I want to see an attempt for this.
Author Response
Response to Reviewer 3 Comments
Point 1: This paper compares the CRP with the laser scanners. Due to the significant improvement in digital cameras, computation powers, and price, it is clear that CRP and other image based scanning methods are going to be really popular.
While this paper provides the comparison of this methods, this paper is not ready for submission yet:
Response 1: We appreciate the Reviewer’s suggestion.
Point 2: The paper is 21 pages with out even one equation. The writer needs to provide in details that how each of these methods work. What are the fundamentals of each methods, and what are the general pros and cons of them, before even diving into the main process. The authors used these methods like a black box while these methods are described in details in lots of researches before. I don't think this paper is publishable with out covering the mathematics, and fundamentals of each methods in details.
Response 2: We appreciate the Reviewer’s suggestion. We thank the anonymous Reviewer for her/his time spent to suggest improvements for our paper. As the Reviewer says, there are many studies on photogrammetry methods. This study is based on an applied approach, and the most significant studies have been mentioned in the paper to not be excessively repetitive.
Point 3: The result of imaged based method can be vastly improve via different lens, even different sensor, and proper calibration. I would like to see the result for this parameters.
Response 3: We appreciate the Reviewer’s suggestion. As mentioned in point 2, the accuracy of photogrammetry depends on many factors, besides those mentioned in the paper, so we agree with the Reviewer’s opinion. However, it would mean to go beyond our goals, particularly when the study specifies that the factors mentioned could meet those goals. We are grateful for the Reviewer’s comments and they will be considered in further research studies.
Point 4: The images based method has multiple steps, such as feature extraction (sift, surf,...), feature matching, epipolar connection, RANSAK,... each of this steps serve as a hyperparameter that can change the final result significantly, I would like to see the result of these parameters.
Response 4: We appreciate the Reviewer’s suggestion. There are many studies on the parameters mentioned by the Reviewer. However, given the structure and goals established, their analysis in this paper would exceed the length and the essential aspects of those goals. Nevertheless, we are grateful for the Reviewer’s comments and they willl be considered in further research studies.
Point 5: Figure 2 is not informative at all.
Response 5: We appreciate the Reviewer’s suggestion. We are grateful for the Reviewer’s contribution, so Figure 2 has been removed.
Point 6: The x limits axis in both plots in figure 10 need to be cropped
Response 6: We appreciate the Reviewer’s suggestion. We proceed to modify figure 10.
Point 7: Figure 11 is not clear at all.
Response 7: We appreciate the Reviewer’s suggestion. We proceed to modify figure 11.
Point 8: The laser scanner also has its own challenges to improve the result, I want to see an attempt for this.
Response 8: We appreciate the Reviewer’s suggestion. We have used the TLS as a starting point to compare the results of the image networks. Going more deeply into the accuracy of equipment or procedures in relation to these remote sensing techniques exceeds the goals of this paper. However, the Reviewer’s opinion is included in the conclusions and it will be considered in further research studies.
Reviewer 4 Report
It is a interesting paper on the well-known topic of validation of image-based modelling techniques in the framework of built heritage documentation (in connection to HBIM modelling). The contents of the paper are efficiently presented with adequate references, however (even though I'm not native English speaker and I don't feel qualified to judge the grammar of the paper) I would strongly suggest a careful revision as regards the language, choice of terms, costruction of some of the sentences that are not very clear (eg. sentence reported in Lines 38-39-40, which seems incomplete). Also, the conclusions maybe should be deepened...
In the following, some points that maybe need to be clarified/typos:
Line 71-72. Not clear, maybe it should be better explained. Even if the final outcome (point clouds) of SfM techniques and TLS can be compared, I do not think it is quite correct to affirm that the work processes of these two techniques are similar.
Line 81. It should be better explained how the representation scale affects on the accuracy of SfM photogrammetry.
Line 142. I think that laser metre and tape metre can be defined as traditional range survey techniques (rather than topographic)...
Line 152. It should be explained with more details why Leica Disto S910 has been used and, more generally, how direct measurements have been used considering the aims of the presented research. In the following paragraphs is not clear to me how these measures are involved.
Line 159. Maybe could be useful a summary table in order to compare principal specifications of the different sensors/lenses used, and the parameters of the camera during the acquisition (eg. resolution, aperture, exposure time, ISO, etc.).
Line 172. How these parameters have been adjusted? A color chart has been used?
Line 178. It would be interesting to consider the metric accuracies observed on Ground Control Points and Control Points during the photogrammetric process.
Line 317. The caption is reported both above (in the text) and belove the Figure 8 (maybe a typo?).
Line 357. Maybe the color scale of the distance analysis performed could be stretched in order to make the discrepancies more detectable since the presence of high values up to 0.94 (maybe due to non significant outliers).
Line 419. Here it looks like the sentence is interrupted.
Line 483. In addition to the number of images used for 3D reconstruction, maybe it should be underlined the importance of the other factors involved in the planning of the acquisition of an adequate dataset (eg. overlapping, acquisition distance, etc)
Author Response
Response to Reviewer 4 Comments
Point 1: It is a interesting paper on the well-known topic of validation of image-based modelling techniques in the framework of built heritage documentation (in connection to HBIM modelling). The contents of the paper are efficiently presented with adequate references, however (even though I'm not native English speaker and I don't feel qualified to judge the grammar of the paper) I would strongly suggest a careful revision as regards the language, choice of terms, costruction of some of the sentences that are not very clear (eg. sentence reported in Lines 38-39-40, which seems incomplete). Also, the conclusions maybe should be deepened..
Response 1: We appreciate the Reviewer’s suggestion. We proceed to modify the text in lines 38-40, as well as the conclusions. The overall text has been reviewed.
Point 2: Line 71-72. Not clear, maybe it should be better explained. Even if the final outcome (point clouds) of SfM techniques and TLS can be compared, I do not think it is quite correct to affirm that the work processes of these two techniques are similar.
Response 2: We appreciate the Reviewer’s suggestion. We proceed to modify the text in lines 71-72.
Point 3: Line 81. It should be better explained how the representation scale affects on the accuracy of SfM photogrammetry.
Response 3: We appreciate the Reviewer’s suggestion. We proceed to modify the text in line 81.
Point 4: Line 142. I think that laser metre and tape metre can be defined as traditional range survey techniques (rather than topographic)...
Response 4: We appreciate the Reviewer’s suggestion. We proceed to modify the text in line 142.
Point 5: Line 152. It should be explained with more details why Leica Disto S910 has been used and, more generally, how direct measurements have been used considering the aims of the presented research. In the following paragraphs is not clear to me how these measures are involved.
Response 5: We appreciate the Reviewer’s suggestion. We proceed to modify the text in line 152.
Point 6: Line 159. Maybe could be useful a summary table in order to compare principal specifications of the different sensors/lenses used, and the parameters of the camera during the acquisition (eg. resolution, aperture, exposure time, ISO, etc.).
Response 6: We appreciate the Reviewer’s suggestion. We proceed to modify the text in line 159.
Point 7: Line 172. How these parameters have been adjusted? A color chart has been used?
Response 7: We appreciate the Reviewer’s suggestion. We proceed to modify the text in line 172.
Point 8: Line 178. It would be interesting to consider the metric accuracies observed on Ground Control Points and Control Points during the photogrammetric process.
Response 8: We appreciate the Reviewer’s suggestion. We proceed to modify the text in line 178.
Point 9: Line 317. The caption is reported both above (in the text) and belove the Figure 8 (maybe a typo?).
Response 9: We appreciate the Reviewer’s suggestion. We proceed to modify the text in line 317.
Point 10: Line 357. Maybe the color scale of the distance analysis performed could be stretched in order to make the discrepancies more detectable since the presence of high values up to 0.94 (maybe due to non significant outliers).
Response 10: We appreciate the Reviewer’s suggestion. We proceed to modify the text in line 357.
Point 11: Line 419. Here it looks like the sentence is interrupted. Line 483. In addition to the number of images used for 3D reconstruction, maybe it should be underlined the importance of the other factors involved in the planning of the acquisition of an adequate dataset (eg. overlapping, acquisition distance, etc)
Response 11: We appreciate the Reviewer’s suggestion. We proceed to modify the text in lines 419 and 483. Two new tables have been included, according to your recommendation.
Round 2
Reviewer 2 Report
Much clearer, thank you to the authors. On a minor point see on line 152, the figure seems to be struck out but it is still referred to in the narrative (see line 148). Perhaps another read through by the authors would be appropriate just to clear up any textual inconsistencies. Otherwise I am happy.
Author Response
We appreciate the Reviewer’s suggestion. We thank the anonymous Reviewer for her/his time spent to suggest improvements for our paper. This aspect has been corrected.
Reviewer 3 Report
While I appreciate the edits, I still think the current form of the paper is not publishable.
The paper should be self-explanatory, so including some details on how these methods are working is necessary. You need to address the fundamentals of both methods.
The current shape of the paper has this magic box approach for the two methods without explaining why we got the current result. Also, in comparing two methods it is a good practice to show the result of both methods are optimized in terms of parameters (locations, lens,....). The current result can be completely anecdotal.
Thanks for the edits in the figures.
Author Response
We appreciate the Reviewer’s suggestion. We thank the anonymous Reviewer for her/his time spent to suggest improvements for our paper. The explanation of the methods has been improved and new references have been added. The reviewer should understand that the paper is focused on applicability and not on the fundamentals of the methods. Likewise, Table 1 shows the parameters of the SfM method where the reader can know the parameters that were used in the study.
