Viewpoint Generation Using Geodesics and Associated Semi-Automated Coverage Path Planning of Panels for Inspection
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsThis paper presents a geometric algorithm for the coverage path planning of panels used for aerospace applications using a generic camera model which can represent area inspection techniques like thermography and laser shearography. However, I would like to provide some feedback on the paper to further enhance its quality:
1. There are some grammar errors in the abstract and paper, and it is necessary to revise and embellish them.
2. In the introduction, there is little comparison of related literatures, please refer to more current latest literatures (General references should be more than 20). For example, you can refer to the following literature:
Wu L, Huang, X.D., Cui J.G., Liu, C., Xiao, W.S. Modified adaptive ant colony optimization algorithm and its application for solving path planning of mobile robot [J]. Expert Systems with Applications, 2023, 215: 119410.
3. It is suggested to add some experiments of well-known improved algorithms to verify the advantages of the proposed algorithm in solving path planning, the effectiveness of this method needs to be further proved.
4. Limitations of the proposed method are not discussed.
Comments on the Quality of English LanguageNO
Author Response
1. There are some grammar errors in the abstract and paper, and it is necessary to revise and embellish them.
Response
Grammer has been checked and improved using online tools.
2. In the introduction, there is little comparison of related literatures, please refer to more current latest literatures (General references should be more than 20). For example, you can refer to the following literature:
Wu L, Huang, X.D., Cui J.G., Liu, C., Xiao, W.S. Modified adaptive ant colony optimization algorithm and its application for solving path planning of mobile robot [J]. Expert Systems with Applications, 2023, 215: 119410.
Response
A few extra references have be added regarding the coverage path planning using drones for coverage path planning, and also another survey paper which mostly covers optimization based algorithms. There is no literature which addresses the specific problem of panel coverage.
3. It is suggested to add some experiments of well-known improved algorithms to verify the advantages of the proposed algorithm in solving path planning, the effectiveness of this method needs to be further proved.
Response
Implementations of other algorithms and comparison with the same are beyond the scope of this paper but may be explored in a future work. The intention behind this paper are not to prove that the method presented is the ‘best possible’, but rather to present the breakdown of the current problem (scanning the panel using camera viewpoints, which has not been addressed in previous literature) and to solve it using a technique that is simple, fast and works for most of the use cases.
The section on computational complexity (Section 5.9) has been expanded to cover this point in detail.
4. Limitations of the proposed method are not discussed.
Response
A new section (Section 5.8) has been added to discuss the limitations of the proposed method.
Reviewer 2 Report
Comments and Suggestions for AuthorsIn this paper authors proposes a geometric algorithm for the coverage path planning planning problem. The focus is made on a real application of aerospace applications panels. As a sensor a generic camera model is implemented.
The paper starts with a well conducted introduction, which includes the motivations as well as the state of art review. The proposed methdo is then well detailed, several results has been also presented and fully commented which is greately appreciated.
Proposed algorithms are well stated and correcetelly presented and commented, which of a great help for readers.
Normally geometric method are recommended for such applications any time shapes and surfaces are known or correcetelly defined, meanwhile optimization and heuristic approaches are recommended for hazardous shaped or merly defined surfaces... Authors method is typically a geomotric method, and authors claim their method to be simple, in regard to NP-hard optimization approches, while they did not assessed the spatio-temporal complexity of their proposition to proof its complexity effectiveness.
Metric comparatives are missing in this paper, a paragraph with metric comparatives toward classical geometric methods is needed to assess the propsal toward similare geometric approches.
A metric comparative is also needed toward optimisation and heuristic approches....
Author Response
Implementations of other algorithms and comparison with the same are beyond the scope of this paper but may be explored in a future work. The intention behind this paper are not to prove that the method presented is the ‘best possible’, but rather to present the breakdown of the current problem (scanning the panel using camera viewpoints, which has not been addressed in previous literature) and to solve it using a technique that is simple, fast and works effectively for most of the use cases.
Computational complexity is discussed in Section 5.9 (although not with rigorous proofs), and also in the conclusion it is noted that the method can be made significantly faster by adopting the half edge data structure, and also Algorithm 6 and 7 can be improved by searching an appropriate subset of nodes instead of iterating over all nodes. However, because the current (‘inefficient’) algorithm still runs in just a few seconds on a normal laptop, the run speed was not considered a significant problem and the focus was more on simplicity rather than the efficiency of the algorithms. This can be improved in future work.
Reviewer 3 Report
Comments and Suggestions for AuthorsComments to the Author
In this paper, a geometric algorithm for coverage path planning of panels using camera viewpoints is proposed, which is for non-destructive testing of panels used in aerospace applications. The description of the methods and analysis of experiments are sufficient. Below are some suggestions for further improvement of the paper.
1. Double check the English language of the whole text. It helps readers to read by avoiding too many long sentences and hard-to-read text, such as the sixth sentence in the abstract. Also, according to some standards and norms of academic writing, the symbol used to refer to specific methods should be an Em dash(—) instead of a Hyphen(-). In the conclusion, the phrase ‘easy to use’ is supposed to utilize '-' to combine into one word.
2. The subtitles ‘(a),(b)’ of figures, such as Fig.2, 3, and 14, should be aligned.
3. It is suggested that the sentence in line 366 be moved before the pseudocode of Algorithm 4. Also,
4. The subtitles—Indirect method and Direct method in the Results section are unnecessary. Sections for test results of each panel shape can simply consist of three parts—analysis of results using indirect method, analysis of results using direct method, and analysis of comparative results.
5. I recommend the authors merge Figure 29. into Figure 28., which would enhance consistency and readability by ensuring every section for test results of each panel shape includes two figures.
6. It is recommended the paper uses the passive voice and avoids excessive first-person words, such as ‘we’.
Comments for author File: 
Comments.pdf
Minor editing of English language required.
Avoid too many subordinate clauses in one sentence.
Author Response
These suggestions have been implemented to the maximum extent possible, except for point 4, as the subtitles of Direct and Indirect method are necessary for clarity to the reader.
Reviewer 4 Report
Comments and Suggestions for AuthorsThe paper introduces an in-house developed coverage path planning algorithm. Despite the considerable efforts invested in this research, the paper lacks scientific merit and novelty.
First of all, the reviewer don't see any advantages of the presented method with respect to the existing ones.
In addition this paper focuses on decomposing the structure and do not show much effort in actual planning of the path.
Results are compared in Section 5 (direct vs indirect) but no valid metrics are used to compare the results.
Lastly, the structure of the paper must be improved since the paper in general, is extremely difficult to follow.
I recommend revising the paper including the title, introduction, etc. such that the paper solely focuses on the decomposition. The proposed methods should clearly state its merits.
Comments on the Quality of English LanguageWritten well in general, but still need to revise the quality of English throughout the paper.
Author Response
=======================
Thank you to the reviewer for their useful feedback.
=======================
Comment:
The paper introduces an in-house developed coverage path planning algorithm. Despite the considerable efforts invested in this research, the paper lacks scientific merit and novelty.
First of all, the reviewer don't see any advantages of the presented method with respect to the existing ones.
Response:
I have added the following paragraph towards the end of section 1:
The method presented in this paper has the advantage over optimization based techniques in being fast and simple to implement. Its advantage over other geometry based techniques is that it is much more versatile than slicing based methods. Finally it has an advantage over just manually planning the path (eg. as presented in our paper Reference 4) in that it is semi-automatic and requires much less human intervention than planning the pose of each individual viewpoint.
Thus, the advantage of the method is clearly presented. Regarding novelty, we believe that there is some novelty related to (a) the use of geodesics for creating
viewpoints, and (b) the approaches used for limiting the propagation of viewpoints. While these innovations may not be drastic, they have lead to a very simple method of solution for the class of problems addressed.
=======================
Comment:
In addition this paper focuses on decomposing the structure and do not show much effort in actual planning of the path.
Response:
One of the major advantages of our approach is that our approach of the decomposition of the structure into a set of regularly arranged viewpoints makes the travel planning trivially simple (as shown in section 6). Additionally, the application for which this algorithm is presented (Non Destructive Testing) is such that the inspection time is much greater than the travel time. The robot can move from one viewpoint to another in a few seconds, while inspection at a viewpoint may take several minutes. Thus, optimizing the viewpoint travel times is not of much importance in this application. This point is explicitly mentioned in Section 6, para 1.
=======================
Comment:
Results are compared in Section 5 (direct vs indirect) but no valid metrics are used to compare the results.
Responses:
This issue was raised by other reviewers too, in previous rounds. We had addressed this in a few sections (eg. Section 5.2.2) by reporting evaluation using some metrics such as the number of nodes to cover the surface. We found that for more complex shapes, the further we get from the central assumption that near rectangular patches are being fitted to cover the surface, the more likely that this technique will not work well (Noted in section 5.8). We are dependent on the human operator to judge the quality of the coverage produced. If the quality is not good enough, the operator can scrap these viewpoints and start again with different start criteria (via the GUI). This is why it is a 'semi-automated' method. This point is also noted in section 5.8. Section 5 focuses on qualitatively understanding the results rather than evaluating metrics.
Another metric that could be used is computational time, which is explained in Section 5.9. It is again mainly dependent on the surface being used (size and number of triangles). In general, because it is very fast, we found that this metric did not reveal significant information.
=======================
Comment:
Lastly, the structure of the paper must be improved since the paper in general, is extremely difficult to follow. I recommend revising the paper including the title,
introduction, etc. such that the paper solely focuses on the decomposition. The proposed methods should clearly state its merits.
Response:
Responding to the comment that the title of the paper could reflect the centrality of the decomposition, we suggest the following new title for the paper:
"Viewpoint Generation Using Geodesics and Associated Semi-Automated Coverage Path Planning of Panels for Inspection" in place of the earlier "Automated Coverage Path Planning for Inspection of Panels using Camera Viewpoints"
I have added the following paragraph at the end of section 1 to make the focus on "decomposition" / "viewpoint generation" more clear:
It is to be noted that the current paper focuses primarily on planning the viewpoints given start criteria provided by the human user, by decomposing the surface into regularly arranged neighbouring camera viewpoints. The user input is very important to achieving proper coverage using this method, thus making it a semi automated technique. Once the set of these viewpoints has been generated, they can be linked together into a path. However, the travel planning process can be done trivially, or by using known methods to solve Travelling Salesman problem (see Section 6). We also note that in the given application of camera based Non-Destructive Testing, the position and orientation of the viewpoints and the coverage of the surface that they provide is of paramount importance, while optimally planning the travel plan or sequence of viewpoints is not that important because the robot can move very fast between viewpoints, but needs to hold at a viewpoint for much longer time to record the data. Hence the algorithms described in this paper are basically about creating the set of viewpoints for covering the surface.
=======================
Round 2
Reviewer 4 Report
Comments and Suggestions for AuthorsAuthors were able to revise most of the comments.
Comments on the Quality of English Languageacceptable
