A Seam Tracking Method Based on an Image Segmentation Deep Convolutional Neural Network

Round 1

Reviewer 1 Report

At the beginning of the review, I want to say that the publication is interesting and, after making some adjustments, should be published.

The detailed comments are as follows:

1. The authors did not clearly define what their overriding research goal is. When reading the introduction, it is difficult to assess whether this is a solution to an existing technological problem during PAW welding, or whether it is the use of a new methodological approach to improve the quality of PAW welding technology through the use of a detailed image segmentation method.

2. In the context of Note 1, there is definitely no description of the technological problem the thesis relates to. Why is the problem of identifying the seam position during welding so important for the quality of the joint? What are the methods currently in use? What is the advantage of the proposed approach? There is information that the advantage is the lack of the need to use auxiliary light source, but there is no justification for the declaration of "improvement the quality of the process".

3. In addition, how do the authors understand "quality" in the PAW process? The described method can, to some extent, be a substitute for visual methods of assessing the quality of a joint, which, however, definitely does not exhaust the issue of the quality of welded joints in general. Concluding this remark, there is an impression that there is no competent description of the research issue in terms of technology.

4. The fourth remark concerns the cited literature. Out of twenty literature, there are nineteen Chinese authors. Rather, it does not describe the current state of this issue in world literature. Therefore, the bibliography definitely needs to be improved.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Overall, the work presented in this paper is commendable, but there are a lot of small issues which have to be addressed.

First of all, the quality of English language is very low so I suggest the authors to ask for a help in correcting that.

The entire paper is full of vague, approximate expressions, like “about 40mm×20mm”, “can be of 0.07 mm”.  One of the most annoying example is (line 105) “the pixel size is about 0.0625mm” !

It is a technical paper, so please be accurate in your expressions.

Lines 21-23: “With the continuous development of science and technology in human society, the traditional offline programming robot welding can no longer meet the increasingly diverse welding needs, the development of welding automation is urgently needed.” – this is a dramatic assumption, which is not backed-up by any proves (literature references). The software packages for offline robot programming (for welding or other robotic operations) have made huge leaps lately. Please reconsider (or back-up with literature references) this assumption.

Figure 2 brings nothing new; it just repeats the information brought by figure 1.

Conclusion section is very short and ends abruptly. Perhaps a graphical depiction of the main results will improve the readability of this section.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf