Robust Registration of Multi-Source Terrain Point Clouds via Region-Aware Adaptive Weighting and Cauchy Residual Control

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

1. In Formula (7), the weight coefficient αk is calculated solely based on the partition average residual  rk , ​​ignoring the variation in residual distribution patterns​​. I think this would undermines the theoretical rigor of the proposed method. Since residuals in complex terrain areas often exhibit high dispersion, reliance on the mean alone fails to distinguish between low-mean–high-variance regions and high-mean–low-variance regions. The authors should address this limitation with further analysis and discussion.

2. Experiments used only a single mining area dataset from Inner Mongolia, failing to validate the proposed method’s generalizability to other landforms.

3. The registration visualization  in Figure 7 ​​lacks annotations​​ (e.g., arrows, bounding boxes, or color-coded circles) to highlight key improvement areas, making it difficult for readers to intuitively discern performance differences between Cauchy-AdaV2 and baseline methods.
​​
4. The analysis ​​omits discussion on the method’s sensitivity to initial positions​​ (e.g., large rotation scenarios). Supplementary experiments testing robustness under significant initial misalignment are recommended.

Author Response

We sincerely thank the reviewers and teachers for their valuable comments and hard work. We have organized our point-by-point responses, please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

This article proposes a dual strategy of regional adaptive weighting and Cauchy residual control for the registration problem in complex terrain, which is quite innovative. Here are some specific suggestions:

1. Figure 3 is not clear and in fact, it fails to convey importantinformation. It is suggested to add a detailed display of local areas and use color images to show the changes in terrain structure.
2. The method for extracting ground points is not explained. In fact, point clouds from different sources have significant differences in the representation of ground objects. If ground points cannot be accurately extracted, the remaining noise will interfere with the algorithm proposed in this paper. Therefore, please provide a detailed explanation of how filtering is performed in the preprocessing stage.
3. The principle behind the settings of minw = 0.4 and maxw = 1.0 in Formula 7 is not explained. Maybe we can conduct a parameter sensitivity experiment to explore this.
4. Figures 7 and 10 still fail to convey information effectively. Maybe it is possible to mark the local magnified areas.
5. The experimental data in this paper does not clearly reflect the particularity of the mining area. It is suggested to supplement data from urban areas with regular structures and natural areas with extensive vegetation coverage to verify the effectiveness of the algorithm. Open datasets can be used.

Author Response

We would like to sincerely thank the reviewers and teachers for their valuable comments and kind guidance. We truly appreciate your hard work. We have carefully organized our responses; please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

Comments and Suggestions for Authors

Please refer to the attachment for details.

Comments for author File: Comments.pdf

Comments on the Quality of English Language

The English language is excellent and requires no improvements.

Author Response

We would like to sincerely thank the reviewers and teachers for their valuable comments and kind guidance. We truly appreciate your hard work. We have carefully organized our responses; please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The additional experiments effectively address the problems raised in the previous version of the manuscript, and I think it is suitable for acceptance.

Reviewer 3 Report

Comments and Suggestions for Authors

The authors have comprehensively and effectively addressed all the revision comments raised in the first round.

Comments on the Quality of English Language

Regarding the English language quality, no problems affecting reading or academic expression were found in the second review. The language is standardized and fluent