Recognition and Tracking of an Underwater Pipeline from Stereo Images during AUV-Based Inspection

Round 1

Reviewer 1 Report

This paper proposed a practical method to detect and track underwater pipelines by using stereo image information, and the authors illustrated the approach with detailed mathematical background. However, there are still some points can be improved in the paper.

Point 1: The data processing pipeline is shown in Figure 4. The letters in the figure don't seem to line up very well making it hard to read. The authors may want to remake a figure with mostly the same content but a better alignment of the letters.

Point 2: There are too many lines starting with the bullet, such as, the algorithm in 4.2.4.1 and the conclusion. The authors may want to make these contents clearer by showing the algorithms in pseudocode style and rearranging the conclusion part. Besides 4.2.4.1 and the conclusion, the authors should carefully check other parts of the paper.

Point 3: There are many ordinal numbers used in the paper, but the numbers in each paragraph seem to continue the previous value instead of restarting from one. Is this a mistake?

Point 4: The distance item in Table 1 starts from 2 to 4, but in Table 2 it starts from 4 to 2, they should be consistent in order.

Point 5: Some of the contents in the introduction section may be a little redundant, the authors may want to focus more on the strongly related works and remove some unnecessary parts.

Minor editing of English language required

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

Although we must appreciate the sincere efforts put forward by the authors in composing the manuscript, the following comments must be addressed by them while revising the paper:

[1] "Abstract": Grammatical error: "Approaches to determining......"

[2] "Related Studies": This sub-section needs recast. It is not written chronologically n the context  of the proposed development. Authors merely described the 'content' of the 'references'., e.g. [6].....then [12] etc. Also authors should comment on how these literature are influencing the present work.

[3] Figure 1: It is not clear how the AUV is positioned with respect to UP dynamically. Mere schematics of 'Camera L' & 'Camera R' will not identify the scientific problems to be solved. The motion of the AUV is very crucial here.

[4] Figure 2 is OK. But what is the significance? What sort of scientific breakthrough in knowledge can be ascertained from this figure?

[5] Figure 3: What will happen when UP is extremely large and beveled (i.e. tapered pipes)? The geometric model of fig. 3 is too simplistic.

[6] Authors must mention the details of the experimental site.

[7] After completing the reading of the paper, it may be concluded that authors have unnecessarily extended a rather 'simple' concept.  The paper must include a few representative relatively complex case-studies wherein challenges for detecting a UP can be debated, such as dim light, entanglement with algae / water plants, dynamically-moving fishes and other aquatic fauna, miscellaneous floating objects etc.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 3 Report

The paper studies the problems of recognition and tracking underwater pipeline (UP) based on AUV by binocular vision technologies. Three techniques for initial recognition of the direction of UP upon its detection has been analyzed. The paper is significance for practical engineering applications.  

Question:

(1) The images used by the experiment are continuous photos or videos? The frame rate of 2 seconds per frame seems very low? Why not use more higher frame rate of videos?

(2) The UP’s boundary seen in the picture is not clearly. More clear image should be used.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 4 Report

A stereo vision-based 2D/3D tracking method for undersea pipe/line inspection is proposed, analyzed and test results reported. The method appears to be simple and reliable in terms of engineering implementation and therefore is significant in terms of cost-effectiveness and practice.

Author Response

Please see the attachment

Author Response File: Author Response.pdf