An Approach for Spectrum Extraction Based on Canny Operator-Enabled Adaptive Edge Extraction and Centroid Localization

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This work presents an automated +1st-order spectrum extraction method for off-axis digital holography in adaptive optics. It is intended to manage low SNR circumstances, multi-region interference, and unusual spectrum forms. The technique allows for precise complex amplitude extraction without intervention by combining multi-stage peak screening, adaptive enhancement, edge detection, morphological processing, and connected-component–based centroid localization. In comparison to traditional rectangular window and fixed-threshold approaches, simulation results consistently demonstrate minimal centroid estimation errors (below 0.245 pixels) and significant gains in wavefront reconstruction accuracy, with residual RMSE reductions of around 89% and 87%. Overall, the work is highly driven, methodologically sound, and backed by quantitative and qualitative validation. Following are some comments.

1. You can significantly improve your work by including at least one fully experimental AO/OADH dataset with known or independently estimated wavefront distortions. Add a short subsection validating the method on real-time AO measurements.
2. Discuss generalization of the model. I could only find the fixed values of parameters. gaussian smoothing, distance treshold etc, 
3. add some explanation of sensitity of performance based on the parameters.
4. Improve introduction , include learning-based spectrum or phase extraction strategies https://doi.org/10.3390/photonics12080810
5. Discuss applications of presented work briefly in the introduction or conclusion https://doi.org/10.1016/j.optlaseng.2024.108553 
6. Results from fig 4 to fig 7 can be further improved. and it is suggested to provide explanations in figs captions.
7. The manuscript focuses on adaptive enhancement and noise mitigation. The following could be used to frame the broader importance of adaptive preprocessing in optical sensing systems. doi: 10.1109/TMC.2023.3312550
8. i would suggest drawing a discussion section describing the strengths and limitations of your study. and projected applications.

 

Author Response

Dear reviewer,

Thanks very much for taking your time to review this manuscript and providing us with the valuable opportunity for revision. We feel great thanks for your professional review work on our paper. According to your valuable suggestions, we have supplemented and revised the corresponding parts in the latest version of the manuscript and highlighted in red for clarity. These comments are all valuable and very helpful for revising and improving our paper. We have carefully responded to all your comments point by point, and the content that could be summarized and incorporated into the latest manuscript has been marked with red underlined text.

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

In this paper, the authors propose a spectrum extraction algorithm based on the Canny operator, combined with adaptive edge extraction and centroid localization, to address the issues of poor adaptability to irregular spots and multi-region interference during the demodulation of OADH interference fringes. Simulation results are provided to demonstrate the feasibility and effectiveness of the proposed method. The topic is relevant and potentially valuable for readers of Photonics. However, several aspects of the manuscript require improvement before it can be considered for publication.

1. In lines 171–172, the authors state: “According to the characteristics of irregular light spots, σ = 5 is set in this paper to achieve the optimal balance between noise suppression and peak preservation.” It is unclear whether the choice of σ = 5 is specific to a particular scenario or experimental setup. How can the optimal value of σ be determined under different application conditions? Similarly, in line 204, the parameter is defined as (d_center=min(M, N)/16). It is not clear whether the value “16” is a fixed constant or an adjustable parameter. Please clarify the rationale behind this choice and discuss whether it should be tuned for different system configurations.
2. In lines 317-318, “The centroid (𝑢̅_{𝑘𝑜𝑝𝑡} ,𝑣̅_{𝑘𝑜𝑝𝑡}) of this region is selected as the precise centroid coordinates of the +1st-order spectrum.” Does this imply that the chosen value corresponds to the largest weight determined by the area size? Is there any mathematical justification supporting this choice? If so, it would be beneficial to include a brief explanation or derivation here to improve clarity and help readers better understand the rationale behind this parameter setting.
3. In line 347-349, “although the rectangular window is indeed located within the +1st-order spectrum spot region, the centroid localization accuracy is compromised by the irregular spot morphology and significant noise interference;”. Compared with the rectangular window, the proposed method indeed demonstrates better performance. However, one key factor affecting the performance is the filter window size, as different window sizes can significantly influence the level of induced noise. It would therefore be helpful to clearly specify the window sizes used for both filtering methods and discuss how the choice of window size impacts the results.
4. In the comparison among the Canny, Rect, and Thresh methods, the manuscript states that the Rect and Thresh approaches “fail to achieve valid wavefront reconstruction.” Under such circumstances, the corresponding RMSE values may no longer reliably reflect reconstruction quality. Therefore, it would be helpful to clarify how RMSE should be interpreted in these failure cases, or to consider introducing additional evaluation metrics that better capture the validity of the reconstruction results.

In summary, I recommend that the manuscript undergo minor revision before it can be reconsidered for publication.

Author Response

Dear Reviewer,

Thanks very much for taking your time to review this manuscript and providing us with the valuable opportunity for revision. We feel great thanks for your professional review work on our paper. According to your valuable suggestions, we have supplemented and revised the corresponding parts in the latest version of the manuscript and highlighted in red for clarity. These comments are all valuable and very helpful for revising and improving our paper. We have carefully responded to all your comments point by point, and the content that could be summarized and incorporated into the latest manuscript has been marked with red underlined text.

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

The manuscript by Li et al. presented a method of centroid localization for the 1^st-order spectrum in off-axis digital holography. The proposed method, validated via simulation, may increase the robustness of wavefront reconstruction for adaptive optics, especially under noise. The method is technically sound. However, before the manuscript can be considered for publication, I have the following comments which I hope the authors could address.

Major points:

1. The term “noise” was very loosely used in the manuscript. For example, line 143 and line 337. What are the possible sources of noise in the system that the authors considered? What is the definition of the signal-to-noise ratio that the authors used? In addition, how the “different types of noise” (line 337) were added to the simulation should be clearly stated in the manuscript.
2. Figure 4: For completeness of the verification, I suggest showing the same with no noise added. This will provide baselines to compare how different filtering methods perform. This will also provide a better sense on how noise affects those filtering methods. In addition, could the author clarify under what condition the noise would be strong?
3. Figure 5: There is a systematic offset for the centroid localization (i.e. the CEE value cannot reach 0). Is it because there is migration of the 0^th-order spectrum even under the no-noise case? Could the authors provide the reason for the cause of this subtle but consistent offset?
4. All the figure captions were too simple. More details of how the figures were obtained should be included in the captions for the readers to follow.

Minor points:

5. Introduction: I suggest adding a more in-depth discussion about the cause of wavefront distortion under atmospheric turbulence to better orient the readers. The atmospheric turbulence would cause fluctuation of refractive index, characterized by refractive index structure parameter [M. Li et al., Investigation on the UAV-To-Satellite Optical Communication Systems, IEEE Journal on Selected Areas in Communications, 2018], which is the key to the wavefront distortion. This further highlights the importance of the use of adaptive optics for real-time application [Y. Guo et al., Direct observation of atmospheric turbulence with a video-rate wide-field wavefront sensor, Nature Photonics, 2024]. Those relevant references should be cited and discussed to broaden the scope of introduction.
6. Line 382: The reference of the “field experimental data” should be included.

Author Response

Dear Editors and reviewers,

Thanks very much for taking your time to review this manuscript and providing us with the valuable opportunity for revision. We feel great thanks for your professional review work on our paper. According to your valuable suggestions, we have supplemented and revised the corresponding parts in the latest version of the manuscript and highlighted in red for clarity. These comments are all valuable and very helpful for revising and improving our paper. We have carefully responded to all your comments point by point, and the content that could be summarized and incorporated into the latest manuscript has been marked with red underlined text.

Please see the attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 3 Report

Comments and Suggestions for Authors

The authors did an excellent job in responding my comments. The manuscript got improved significantly. Moreover, the authors further included experimental verification, which is even more satisfying. I think the revise manuscript is worth being published. I only have a few comments in terms of the formatting.

1. The authors shall give the expression of the signal-to-noise ratio (SNR) at its first appearance after the introduction (line 187 instead of line 371). In addition, the notation should be written clearly as the ones in the response letter.
2. Reference [2] and [3] in the revised manuscript are the same. Also, reference [13] and [14] are the same. The style of the citations should be consistent and adapt to the journal style.

Author Response

Dear Reviewer,

We sincerely thank you for your meticulous review of our work and the valuable suggestions you put forward on formatting details, and we are especially grateful for your recognition and approval of our efforts in the previous manuscript revisions! We also appreciate the precious time you have devoted to reviewing this manuscript and the highly valuable revision opportunities you have provided us with, and we extend our sincere gratitude for your professional review of this paper. All the comments you raised are both targeted and instructive, and have played a pivotal role in our revising and refining the paper as well as enhancing the overall quality of the manuscript.

We have strictly followed your valuable suggestions and completed the supplementation and revision of the corresponding content in the updated version of the manuscript. To facilitate your clear review of the revised content, all revised parts have been highlighted in red; in particular, the core content that can be directly summarized and incorporated into the manuscript has been additionally marked with red underlining for special attention. In response to all the comments you raised, we have provided careful and detailed point-by-point responses, and all the relevant revised and supplementary content has been fully incorporated into this revised manuscript.

Please see the attachment.

Author Response File: Author Response.pdf