A 3D Vision-Based Weld Seam Extraction Method for Arbitrarily Placed Spatial Planar Weldments

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The article titled "A 3D Vision-based Weld Seam Extraction Method for Arbitrarily Placed Spatial Planar Weldments" makes a significant contribution to the field of automated welding. The proposed methodology, which combines the use of a 3D camera with an improved RANSAC algorithm, presents an innovative and effective approach for weld seam extraction in workpieces with complex geometry and arbitrary positioning.

The experimental results presented are solid, demonstrating the accuracy and applicability of the method in practical industrial production scenarios. The structure of the article is clear, and the proposed solutions efficiently address the challenges identified in the existing literature.

Given the relevance of the topic, the quality of the methodology, and the results presented, I approve the publication of this article.

Author Response

Comments 1: The experimental results presented are solid, demonstrating the accuracy and applicability of the method in practical industrial production scenarios. The structure of the article is clear, and the proposed solutions efficiently address the challenges identified in the existing literature.

Response 1: Thank you for your approval of this research.

Reviewer 2 Report

Comments and Suggestions for Authors

Here are some suggestions for improving the article, which could enhance its impact and accessibility for a broader audience:

1. Clarity in Mathematical Exposition:

• Section 3.2.4, which describes the "Optimizations for Final Output" and the associated mathematical formulas, could be simplified or explained more clearly. For example, adding diagrams or illustrative examples would help clarify the optimization process.

2. Contextualization of Results:

• A more extensive discussion on how the proposed method compares to existing ones would be beneficial. Direct comparisons with other techniques could provide a clearer perspective on the innovative contribution of this work.

3. Expansion of the Conclusion Section:

• The conclusions could benefit from a more detailed discussion of potential practical applications of the method, as well as identifying possible future research directions.

4. Grammatical and Stylistic Accuracy:

• While the article is well-written, there is room for minor improvements in sentence formulation and structure to ensure clarity and coherence.

These enhancements could help make the work more accessible and impactful, particularly for readers who may not be experts in the specific technical aspects discussed.

Author Response

Comments 1. Clarity in Mathematical Exposition:

Section 3.2.4, which describes the "Optimizations for Final Output" and the associated mathematical formulas, could be simplified or explained more clearly. For example, adding diagrams or illustrative examples would help clarify the optimization process.

Response 1:

Agree. Here are the modifications:

• An example that illustrates the effect of Rule 5 on the segmentation results is added. （264~271）
• Clearer statement of the implementation steps is used. （253~263）

Comments 2. Contextualization of Results:

A more extensive discussion on how the proposed method compares to existing ones would be beneficial. Direct comparisons with other techniques could provide a clearer perspective on the innovative contribution of this work.

Response 2:

Agree. A more substantial discussion is presented now. (532~552)

Comments 3. Expansion of the Conclusion Section:

The conclusions could benefit from a more detailed discussion of potential practical applications of the method, as well as identifying possible future research directions.

Response 3:

Agree. Added possible application scenarios, current limitations and future priorities in the conclusion section. (559~588)

Comments 4. Grammatical and Stylistic Accuracy:

While the article is well-written, there is room for minor improvements in sentence formulation and structure to ensure clarity and coherence.

Response 4: 

Thanks for pointing out. We have thoroughly checked the article for grammar and punctuation, correcting multiple errors (for ease of reading, these corrections are not highlighted in the text.)

Reviewer 3 Report

Comments and Suggestions for Authors

Review of the paper:

A 3D Vision-based Weld Seam Extraction Method for Arbitrarily Placed Spatial Planar Weldments

Authors:

Bo Yang, Zhengtuo Wang, Yuetong Xu, Songyu Hu and Jianzhong Fu

The overall paper structure is commendable, adhering to the accepted format for scientific papers and logically presenting the content. The findings and results are likely to pique the interest of future Readers. The subject matter presented in this paper is of great interest from the perspective of future applied presented methodology that identifies weld seams from arbitrarily placed spatial planar weldments in industrial practice. Nevertheless, some adjustments and additions are necessary to meet the publication standards of the Applied Sciences—MDPI Journal. The following criticisms are noted:

·        In the Abstract section, please do not use abbreviations that the authors will only translate in the main text. There is space for this later in the article. Related to this comment is the suggestion to read the whole text to decipher the abbreviations.

·        Line 87: When describing the structure of the article, the authors forgot to mention what is in Section 1.

·        Figure 3: Please add a legend to the caption of the figure explaining each of the cases analyzed.

·        Figure 5: Include in the caption of the figure how the images are differentiated.

·        Section 5.1. The test bench should be carefully described, defining the components concerned specifically as to model and manufacturer.

·        Table 5: additionally include in the table the percentages relative to the known previous methods, so that the results obtained can be more easily interpreted.

·        Line 511 should be written 3D, not 3D..

 

·        In the Conclusion section, highlight the limitations of the developed methodology, especially if the authors identify potential future areas of application and suggest directions for further research. This provides a comprehensive perspective on the scope and possible challenges of the proposed approach. Is the presented methodology also suitable for welding techniques of very thin and flaccid sheets?

Author Response

Comments 1:

In the Abstract section, please do not use abbreviations that the authors will only translate in the main text. There is space for this later in the article. Related to this comment is the suggestion to read the whole text to decipher the abbreviations.

Response 1:

Agree. The abbreviation “WSE” is removed. (Line 18)

Comments 2:

When describing the structure of the article, the authors forgot to mention what is in Section 1.

Response 2:

Agree. I added a summarization for Section 1 in the last paragraph. (112~113)

Comments 3:

Figure 3: Please add a legend to the caption of the figure explaining each of the cases analyzed.

Response 3:

Agree. A summarization of the four cases is added in the caption of Figure 3. (198~199)

Comments 4:

Figure 5: Include in the caption of the figure how the images are differentiated.

Response 5:

Agree. A brief description is added to Figure 6 (the original Figure 5.) (Line 282)

Comments 5:

Section 5.1. The test bench should be carefully described, defining the components concerned specifically as to model and manufacturer.

Response 5:

Agree. The main parameters of the 3D camera and the turntable are added. (409~412)

Comments 6:

Table 5: additionally include in the table the percentages relative to the known previous methods, so that the results obtained can be more easily interpreted.

Response 6:

Agree. Relative comparisons against the conventional algorithm are added in Table 5, 6, and 7.

Comments 7:

Line 511 should be written 3D, not 3D..

Response 7:

Agree. The whole conclusion part has been modified. This problem does not exist anymore. 

Comments 8:

In the Conclusion section, highlight the limitations of the developed methodology, especially if the authors identify potential future areas of application and suggest directions for further research. This provides a comprehensive perspective on the scope and possible challenges of the proposed approach. Is the presented methodology also suitable for welding techniques of very thin and flaccid sheets?

Response 8:

Thank you for pointing out. This is added in the conclusion now: “Due to the poor performance of RANSAC (even the improved version proposed in the paper) in identifying narrow and long planes, this method only applies to plate structures with a thickness of more than 2 mm.” The conclusion part is also extended now. (573~588)

 

Reviewer 4 Report

Comments and Suggestions for Authors

The manuscript presents a method for vision-based weld seam extraction (WSE) in robotic welding, particularly for arbitrarily placed spatial planar weldments.

In the Introduction, the authors mention that existing studies assume the position and orientation of weldments, are known in advance, and only focus on a single frame of point cloud data for weld seam extraction.

Consider explicitly stating the gap in the existing literature that this study aims to address.

The introduction should include more significant references related to the main problem. Additionally, information on previous work should be listed and discussed to identify knowledge gaps. The introduction lacks information about numerical results, which are necessary to state this study's baseline effectively.

The authors proposed an improvement of the vanilla RANSAC algorithm. Still, they failed to present previous research related to specific improvements on the same algorithm and how they differ from the present research.

Information in Section 2 is too short to constitute a complete section.

Include the reference axis system in Fig 1.

Please include references for the selected p=0.99 in Eq. 6.

Lines 172-173 are confusing; please explain better.

In the title of Fig 3, briefly explain cases a,b,c, and d.

Section 3.4 is practically empty; reconsider organizing the section distribution and content.

In line 305, the authors mention “written in Python style. " Please explain; it was coded in Phyton or in which language?

This phrase in line 377 is incomplete: “the two planes forming edge 𝑝1𝑝2 are coplanar”.

Please explain in Section 5 more details about the calibration procedures, how they are done, and what parameters are modified by this process.

In Table 6, it is reported that the vanilla method correctly segmented 15 out of 21 planes, indicating 6 incorrectly segmented planes. However, the same table also shows that 10 planes were incorrectly segmented using this method. This discrepancy is also observed with the non-clustering RANSAC method. The same issue is shown in Table 7. Please discuss and explain this conflicting information in the section where the results are presented.

The discussion section of the manuscript is too brief, suggesting the need to reconsider the order and sections for clarity.

 

The conclusion is clearly presented and based on the results. The authors also include relevant future work and state the importance of this research in the context of the existing state of the art.

Comments on the Quality of English Language

 The quality of the English language in the paper is generally adequate.

Some sentences are complex and need to be simplified to improve readability.

 

There are minor grammatical issues, such as inconsistent use of articles and some punctuation errors.

Author Response

Comments 1:

In the Introduction, the authors mention that existing studies assume the position and orientation of weldments, are known in advance, and only focus on a single frame of point cloud data for weld seam extraction. Consider explicitly stating the gap in the existing literature that this study aims to address.

The introduction should include more significant references related to the main problem. Additionally, information on previous work should be listed and discussed to identify knowledge gaps. The introduction lacks information about numerical results, which are necessary to state this study's baseline effectively.

The authors proposed an improvement of the vanilla RANSAC algorithm. Still, they failed to present previous research related to specific improvements on the same algorithm and how they differ from the present research.

Response 1:

Thank you for pointing out. Here are the modifications: 

 

• References are updated to more clearly stating the gap in the literature and this study.
• This paragraph is added to explane the lack of a more unified comparison: "Due to the difference in weldment sizes and types, equipment types and accuracy levels, and error evaluation methods in the existing literature, we cannot create a unified benchmark for various methods for comparison. However, the process in this paper successfully solves the problem of WSE in a scenario that is more complex than similar articles and greatly improves the accuracy compared to traditional feature extraction methods. It provides an effective reference for applying robot automatic welding of complex planar welded parts without precise clamping."

For ease of reading, the modified content is not highlighted since almost the whole Section 1 is modified. (27~111)

 

Comments 2:

Information in Section 2 is too short to constitute a complete section.

Response 2:

Agree. The content in original Section 2, including Figure 1, has been moved to the original Section 3.1 to make the article structure more reasonable. (121~126)

Comments 3:

Include the reference axis system in Fig 1.

Response 3:

Agree. Reference axes of the turntable system and camera system are added in Figure 1. (Line 124)

Comments 4:

Please include references for the selected p=0.99 in Eq. 6.

Response 4:

Agree. The reference has been added. (Line 164)

Comments 5:

Lines 172-173 are confusing; please explain better.

Response 5:

Agree. Clearer description is used now. (181~183)

Comments 6:

In the title of Fig 3, briefly explain cases a,b,c, and d.

Response 6:

Agree. A summarization of the four cases are added. (Line 198)

Comments 7:

Section 3.4 is practically empty; reconsider organizing the section distribution and content.

Response 7: 

Agree. Section distribution is corrected (by modifying levels of the section headers.) (Lines 318 and 331)

 

Comments 8:

In line 305, the authors mention “written in Python style. " Please explain; it was coded in Phyton or in which language?

Response 8:

The algorithm implementation follows the syntax of Python, but it is pseudocode after all, so it is called Python style rather than Python language.

Comments 9:

This phrase in line 377 is incomplete: “the two planes forming edge ?1?2 are coplanar”.

Response 9:

Agree. Grammar errors are fixed now. (391~393)

Comments 10:

Please explain in Section 5 more details about the calibration procedures, how they are done, and what parameters are modified by this process.

Response 10:

Sure. Detailed procedures of the calibration are added (four steps). (416~435)

Comments 11:

In Table 6, it is reported that the vanilla method correctly segmented 15 out of 21 planes, indicating 6 incorrectly segmented planes. However, the same table also shows that 10 planes were incorrectly segmented using this method. This discrepancy is also observed with the non-clustering RANSAC method. The same issue is shown in Table 7. Please discuss and explain this conflicting information in the section where the results are presented.

Response 11:

Sure. We added explanation to this phenomenon in the paragraph before Table 6. (484~492)

Comments 12:

The discussion section of the manuscript is too brief, suggesting the need to reconsider the order and sections for clarity.

Response 12:

Agree. A more substantial discussion is presented now. (532~552)

Comments 13:

The quality of the English language in the paper is generally adequate. Some sentences are complex and need to be simplified to improve readability. There are minor grammatical issues, such as inconsistent use of articles and some punctuation errors.

Response 13:

Thanks for pointing out. We have thoroughly checked the article for grammar and punctuation, correcting multiple errors (for ease of reading, these corrections are not highlighted in the text.)

Reviewer 5 Report

Comments and Suggestions for Authors

Dear Authors,

the manuscript is interesting and generally well written. Please follow the comments.

1. Please clearly state the novelty of your research in the abstract. 3D vision is well known techniqe. The only novelty seen in this paper is improved RANSAC algorithm. Please explain the improvements of the algorithm.

2. Introduction. One of the factors influencing the quality of assembly joints, including welded ones consisting of several parts, is the appropriate sequence and order of assembly. Please cite: https://doi.org/10.3390/sym14051013

3. Please provide the model of the 3D camera and parameters (for example accuracy)

4. Table 5, 6, 7 - two decimal places are completely sufficient. More is on the border of measurement error.

5. Conclusions - conclusion (2) - please provide numeral values of improvement results of the proposed algorithm

 

Author Response

Comments 1:

Please clearly state the novelty of your research in the abstract. 3D vision is well known techniqe. The only novelty seen in this paper is improved RANSAC algorithm. Please explain the improvements of the algorithm.

Response 1:

Agree. We have added numerical improvements of the proposed method over traditional methods now. (19~22)

Comments 2:

Introduction:  One of the factors influencing the quality of assembly joints, including welded ones consisting of several parts, is the appropriate sequence and order of assembly. Please cite: https://doi.org/10.3390/sym14051013

Response 2:

Done. (59~61)

Comments 3:

Please provide the model of the 3D camera and parameters (for example accuracy)

Response 3:

Agree. The main parameters of the 3D camera are added. The model of the camera can be found in Figure 1. To avoid redundancy, it is not provided here. (409~412)

Comments 4:

Table 5, 6, 7 - two decimal places are completely sufficient. More is on the border of measurement error.

Response 4:

Agree. All decimals are modified with reasonable numbers of decimal places. (For ease of reading, no hightlight is made)

Comments 5:

Conclusions - conclusion (2) - please provide numeral values of improvement results of the proposed algorithm

Reponse 5:

Agree. Numeral values of improvements are provided both in Table 5, 6, 7 and in the conclusion part. (564~567)

Round 2

Reviewer 4 Report

Comments and Suggestions for Authors

The authors have responded in a detailed and satisfactory manner to each of the suggestions and comments provided during the review process.

The article presents a sound and relevant approach to the development of automated processes and industrial applications, making it a valuable contribution to the field. In addition, the clarity of the methodology allows for easy replication of the experiments and ensures that the results can be rigorously validated by other researchers.

The article's conclusions are supported and align consistently with the results obtained throughout the study. The authors synthesize the findings appropriately.

However, it is suggested that the authors expand their discussion of possible limitations of the study and future scenarios in which their research could be applied or improved.