PriKMet: Prior-Guided Pointer Meter Reading for Automated Substation Inspections

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript addresses a critical challenge in the modernization of power infrastructure monitoring—automated pointer meter reading in substation environments. The authors propose PriKMet (Prior-Guided Pointer Meter Reader), a deep learning-based framework designed to overcome this challenge through a unified hierarchical architecture, route-aware perception fusion, and adaptive offset correction for UAV inspections. The study is well-grounded in both technical rigor and practical relevance, showing promising results on a sizable real-world dataset. Here are some comments:

1. The introduction could cover the literature review in total, instead of having two separate sections, introduction and related work. In this case, the authors need to shrink the introduction part.
2. Some abbreviation definitions are not mentioned, like the UAV, RCNN,SGD…… etc
3. The referencing part needs review, for example, the authors mention that reference [13] is for Liu et al, however, it is different in the reference. The same for ref [17] and [19]. Maybe there are others.
4. Figure 3 is not explained in a sufficient way. what do the levels refer to? What is C2F?
5. The link between Fig.3 and Fig.4 is not clear. As mentioned, Figure 4 should be part of Figure 3, but it is not clear where. Where can we find the Bottleneck module in Fig.3?
6. The label for Figure 3 is very long.
7. Equations 2 and 3 are written as code, not in a mathematical way. In addition, the equation of angles α and .
8. Starting from line 207 till the end of the paper, the equations are not numbered.
9. How can the system detect the reading if the meter glass is not clean or wet? How could the system be robust against the hard industrial environment? The authors didn’t speak about these situations. It should be a certain index for what the results will be in various environmental scenarios.
10. The authors didn’t mention anything related to the hardware of the UAV used and how it will affect the results. What is the affect of the resolution of the camera used?
11. What are Aps and AP50?
12. In section 4.3.3, it is not clear what the interference source is used here. Consequently, what is the interference time?
13. Table 5, What is the vanilla?

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Mandatory statements like Author Contributions are missing! Please insert them according to the MDPI requirements.

All equations/relations must be numbered.

More relevant experimental data is mandatory. Photos taken under different light conditions or weather conditions must also be considered.

Why does the test set in Table 1 contain only 200 images?

Is there any limitation for the meters (in terms of geometry, size)?

Does the software provide error reports, too?

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

Very nice work, and it could be accepted for publication after minor revisions.

1) More details for data acquisition should provided.

2）The equation number should be checked in Pages 7-8.

3) The relationship between Figure 2 and Figure 5 should be provided.

4) The data set is public or not?

5) The value for some indexes of the proposed network should be given in Abstract and Conclusion Sections.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

No other comments