Construction and Visualization of Levels of Detail for High-Resolution LiDAR-Derived Digital Outcrop Models

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This study addresses the challenges of large data volume and performance bottlenecks in the visualization of high-resolution LiDAR-derived digital outcrop models. The authors propose an automated LOD construction and visualization method designed to preserve geological features. The research demonstrates strong practical value, particularly for multi-scale data organization and efficient rendering of elongated cliff-face outcrop models in geological applications. The overall structure of the paper is clear and logically coherent, with well-defined research objectives and a sound technical approach. The following suggestions are provided for further improvement:

（1）Highlight the innovation more explicitly. For example, in the Abstract, the authors could more clearly specify the main improvements of this method compared with conventional approaches. In addition, it is recommended to add a quantitative result (such as the increase in rendering frame rate) after the sentence “Results demonstrate that ...” to enhance the persuasiveness of the findings.

（2）Refine language and formatting details. Some sentences are relatively long; appropriate segmentation or the use of semicolons is suggested to improve readability.

（3）Strengthen the logical connections in the literature review. Although the Introduction and Related Work sections provide a generally comprehensive and systematic overview, the authors should better emphasize the specific research problem and target addressed in this study, and improve the logical transitions between sections. For instance, in Section 2.2, after the opening sentence “The concept of Level of Detail (LOD) was first introduced by Clark (1976)...”, a linking sentence such as “Although originally developed for computer graphics, the concept has since become central to geospatial data management.” could better highlight the geological application context of the study. Additionally, these literatures may be highly relevant to your research: A photogrammetric approach for quantifying the evolution of rock joint void geometry under varying contact states; A novel method for determining the three-dimensional roughness of rock joints based on profile slices; Determination of the minimum number of specimens required for laboratory testing of the shear strength of rock joints.

（4）Add a summary analysis of performance improvements in the Results section. For example, at the end of Section 4.2, it would be beneficial to include a short summary quantifying the overall increase in frame rate after LOD construction, rather than only presenting the raw performance data.

（5）Enhance the conclusion with contextual linkage. It is suggested to add a brief introductory sentence at the beginning of the Conclusion section to restate the research background or the primary challenge, thereby forming a logical closure between the problem statement and the contributions of the work.

Comments on the Quality of English Language

Refine language and formatting details. Some sentences are relatively long; appropriate segmentation or the use of semicolons is suggested to improve readability.

Author Response

Comments 1: Highlight the innovation more explicitly. For example, in the Abstract, the authors could more clearly specify the main improvements of this method compared with conventional approaches. In addition, it is recommended to add a quantitative result (such as the increase in rendering frame rate) after the sentence “Results demonstrate that ...” to enhance the persuasiveness of the findings.

Response 1:

Thank you for your valuable suggestions. We have explicitly highlighted the core contributions, such as the “the feature-preserving mesh simplification algorithm that incorporates vertex sharpness constraint and boundary freezing strategy to retain critical geological features " in line with the reviewers' comments. And we have added a quantitative result after the sentence “Results demonstrate that ...” to enhance the persuasiveness of the findings. The revised Abstract is as follows:

Abstract: High-resolution LiDAR-derived three-dimensional (3D) digital outcrop models are crucial for detailed geological analysis. However, their massive data volumes often exceed the rendering and memory capacities of standard computer systems, posing significant visualization challenges. Although Level of Detail (LOD) techniques are well-established in Geographic Information System (GIS) and computer graphics, they still require customized design to address the unique characteristics of geological outcrops. This paper introduces an automated method for constructing and visualizing LOD models specifically tailored to high-resolution LiDAR outcrops. The workflow begins with segmenting the single-body model based on texture coverage, followed by building an adaptive LOD tile pyramid for each segment using a pseudo-quadtree approach. The proposed LOD construction method incorporates several innovative components: segmentation based on texture coverage, an adaptive LOD tile pyramid using a pseudo-quadtree, and a feature-preserving mesh simplification algorithm that includes vertex sharpness constraint and boundary freezing strategy to maintain critical geological features. For visualization, a dynamic multi-scale loading and rendering mechanism is implemented using an LOD index with the OpenSceneGraph (OSG) engine. Results demonstrate that the proposed method effectively addresses the bottleneck of rendering massive outcrop models.  The models loading time and average memory usage were reduced by more than 90%, while the average display frame rate reached around 60 FPS. It enables smooth, interactive visualization and provides a robust foundation for multi-scale geological interpretation.

Comments 2: Refine language and formatting details. Some sentences are relatively long; appropriate segmentation or the use of semicolons is suggested to improve readability.

Response 2:

We appreciate your valuable feedback regarding the sentence length and readability. In response to your suggestion, we have carefully revised the manuscript to improve clarity and readability by appropriately segmenting longer sentences and using semicolons where necessary. This revision aims to enhance the flow of the text and ensure that key points are conveyed more clearly to the reader. For example:

In Section 4.3

Original:

The results of geometric and texture errors in model simplification are shown in Figure 16. In terms of geometric error testing, the geometric errors of all three simplification methods gradually increase with higher simplification rates. When the simplification rate exceeds 45-55%, the geometric error of our algorithm becomes slightly larger than the other two methods. This increased geometric error may be attributed to our algorithm's consideration of more local detail features and texture characteristics, leading to a slight increase in global integration error. Regarding texture error testing, the texture errors of all three methods also increase with higher simplification rates, but our algorithm achieves the smallest texture error, while the QEM algorithm exhibits the largest. This demonstrates the effectiveness of the constraints and strategies designed in our algorithm for texture protection.

Revised:

We compared the visual effects of different simplification results, taking the simplified model with a 0.3 simplification ratio derived from experiments as an example. As shown in Figure 17, conventional QEM and MeshLab simplifications cause noticeable texture stretching and boundary cracks, whereas our method maintains coherent textures and intact geological features.

 

Comments 3: Strengthen the logical connections in the literature review. Although the Introduction and Related Work sections provide a generally comprehensive and systematic overview, the authors should better emphasize the specific research problem and target addressed in this study, and improve the logical transitions between sections. For instance, in Section 2.2, after the opening sentence “The concept of Level of Detail (LOD) was first introduced by Clark (1976)...”, a linking sentence such as “Although originally developed for computer graphics, the concept has since become central to geospatial data management.” could better highlight the geological application context of the study. Additionally, these literatures may be highly relevant to your research: A photogrammetric approach for quantifying the evolution of rock joint void geometry under varying contact states; A novel method for determining the three-dimensional roughness of rock joints based on profile slices; Determination of the minimum number of specimens required for laboratory testing of the shear strength of rock joints.

Response 3:

We appreciate the reviewer’s insightful suggestions for improving the logical flow of the literature review. In response to the recommendation to strengthen the logical connections, we have revised the manuscript to emphasize the specific research problem and the target of this study more clearly. We have also added a linking sentence in Section 2. Additionally, we have incorporated the suggested references that are highly relevant to the geological context of our research. These additions enrich the background and provide a deeper understanding of the application of LOD techniques in geological research. We are grateful for the highly relevant references suggested by the reviewer. We have now integrated them into the revised manuscript (in the Introduction and Section 2) to strengthen our argument.

Thank you again for your valuable feedback, which has significantly strengthened the manuscript.

Comments 4: Add a summary analysis of performance improvements in the Results section. For example, at the end of Section 4.2, it would be beneficial to include a short summary quantifying the overall increase in frame rate after LOD construction, rather than only presenting the raw performance data.

Response 4:

Thank you for the valuable suggestion. We have added a summary analysis of the performance improvements in the Results section, specifically at the end of Section 4.2. This addition quantifies the overall increase in frame rate after the LOD construction, providing a clearer perspective on the performance enhancements beyond the raw data. The revised section now includes the following summary: Overall, the model's loading time was reduced by more than 95%, while the average display frame rate can reach around 60 FPS.

 

Comments 5: Enhance the conclusion with contextual linkage. It is suggested to add a brief introductory sentence at the beginning of the Conclusion section to restate the research background or the primary challenge, thereby forming a logical closure between the problem statement and the contributions of the work.

Response 5:

We fully agree with the recommendation. To address this, we have added a concise summary at the beginning of the Conclusion to highlight the core challenge of the research and our main contributions.  The revised Conclusion as follows:

This paper has presented a comprehensive LOD-based solution for the efficient visualization of large-scale, high-resolution LiDAR digital outcrop models. Addressing the rendering bottleneck caused by massive data volumes and complex geometries, the proposed approach integrates model segmentation, adaptive tiling, feature-preserving simplification, and dynamic visualization into a unified workflow.

First, we designed a tailored LOD construction pipeline that integrates model segmentation, pseudo-quadtree tiling, feature-preserving simplification, and texture reconstruction. This workflow effectively addresses the unique challenges posed by the elongated and vertical nature of digital outcrops, overcoming the limitations of conventional LOD methods designed for planar or large-area models.

Second, we developed and integrated key algorithmic optimizations. The incorporation of boundary freezing and fallback strategies into the feature-preserving QEM algorithm ensured the protection of both geometric integrity at tile borders and critical geological features during aggressive mesh simplification.

Finally, we established an efficient data organization and dynamic visualization framework. By employing an LOD index and OSG PagedLOD mechanism, we achieved real-time, view-dependent loading and seamless multi-scale rendering of high-resolution LiDAR-derived digital outcrop models on standard computer hardware.

Experimental results demonstrate that the proposed method achieves substantial reductions in loading time and memory usage while maintaining high visual fidelity. Overall, this research provides a practical and scalable solution for 3D geological model visualization, paving the way for advanced applications such as semantic attribute integration and web-based collaborative geological interpretation. Future work will focus on enhancing the adaptability, computational efficiency, and cross-domain applicability of the framework.

Author Response File: Author Response.docx

Reviewer 2 Report

Comments and Suggestions for Authors

This manuscript address the limitation of geological survey from excessive economic, weather condition, remote area, and dangerous zone from fracture and fissues of weak rocks, which utilized with LiDAR to digital outcrop model. Moreover, this study demonstrate the how to reduce the performance of computation and this approach can be reproduced. However, there are some minor comments for improving manuscript. Therefore, I recommend to accept with minor revision.

1. Please revise the abstract, which has to focus and highlight the original of research.
2. Please check the abbreviations in manuscript, I found many mistaken abbreviations. Example is Line 13 "GIS", which mean "Geographic Information System".
3. Section 2: Related Work. It should be literautre research, Related Research, other word, that is better than it. Or the authors can move this section into methodology section.

Author Response

Comments 1: Please revise the abstract, which has to focus and highlight the original of research.

Response 1:

Thank you for your valuable suggestions. We have explicitly highlighted the core contributions, such as the “the feature-preserving mesh simplification algorithm that incorporates vertex sharpness constraint and boundary freezing strategy to retain critical geological features " in line with the reviewers' comments. The revised Abstract as follows:

Abstract: High-resolution LiDAR-derived 3D digital outcrop models are essential for detailed ge-ological analysis. However, their massive data volumes often exceed the rendering and memory capacities of standard computer systems, posing significant visualization challenges. Although Level of Detail (LOD) techniques are well-established in Geographic Information System (GIS) and computer graphics; they are not directly suitable for the elongated, cliff-face morphology of typical outcrops. This paper presents an automated method for constructing and visualizing LOD models specifically tailored to high-resolution LiDAR outcrops. The workflow begins by segmenting the single-body model based on texture coverage, followed by building an adaptive LOD tile pyramid for each segment using a pseudo-quadtree approach. The proposed LOD construction method integrates several innovative components: segmentation based on texture coverage, an adaptive LOD tile pyramid using a pseudo-quadtree, and a feature-preserving mesh simplification algorithm that incorporates vertex sharpness constraint and boundary freezing strategy to retain critical geological features. For visualization, a dynamic multi-scale loading and rendering mechanism is implemented using an LOD index with the OpenSceneGraph (OSG) engine. Results demonstrate that the proposed method effectively over-comes the bottleneck of rendering massive outcrop models. The models loading time and average memory usage were reduced by more than 90%, while the average display frame rate can reach around 60 FPS. It enables smooth, interactive visualization and provides a robust foundation for multi-scale geological interpretation.

Comments 2: Please check the abbreviations in manuscript, I found many mistaken abbreviations. Example is Line 13 "GIS", which mean "Geographic Information System".

Response 2:

Thank you for pointing that out. I have reviewed the manuscript and corrected the abbreviations. The term "GIS" in line 13 has been revised to "Geographic Information Science" as intended. I have also addressed other abbreviation errors throughout the manuscript.

 

Comments 3: Section 2: Related Work. It should be literautre research, Related Research, other word, that is better than it. Or the authors can move this section into methodology section.

Response 3:

We appreciate the reviewer's valuable suggestion regarding the organization of the literature review. We have renamed "Section 2: Related Work" to "Section 2: Literature Review" to more accurately reflect its content. Furthermore, we have strengthened this section by not only summarizing previous studies but also more clearly articulating the research gap that our work aims to fill, which provides a smoother transition to our proposed methodology.

Author Response File: Author Response.docx

Reviewer 3 Report

Comments and Suggestions for Authors

Dear authors and editor,

 

I read with interest the paper “Construction and Visualization of Levels of Detail for High resolution LiDAR derived Digital Outcrop Models”. It is in general a well written paper, but I have envisioned two potential issues. One regards the subject of the paper, that maybe is more focused on software than on remote sensing perspective, and maybe it could fit more on a journal dedicated to software in geosciences. The second one is that in my viewpoint the authors should mention the possibility to include in the pipeline (maybe in future studies) the representation of surface texture/roughness and not only of image texture. The availability of dense lidar points clouds permit to describe very well surface roughness, that can be very useful in the visual interpretation of outcrops. So why not use this information?

Then from a technical perspective I have the following comments:

-figure 10b it is unclear what the different labelled parts represent

- section 4.3 geometric error and texture error. In regard to the geometric error it is unclear to me to which polygons the Hausdorff distance is applied. Then, I’m wondering if the Chamfer distance could be an alternative. In regard to the texture error I think that the image structural similarity index could provide a much more complete error metric. Image and surface texture have a spatial structure that cannot be captured by just RMS of pixels differences.

 

 

 

Author Response

Comments 1: figure 10b it is unclear what the different labelled parts represent.

Response 1:

We appreciate the reviewer's feedback regarding the clarity of Figure 10b. In response, the figure caption has been updated accordingly to reference all these Labels. We hope the revised figure meets the required standard of clarity. The revised figure caption as follows:

Figure 10. Geometric illustration of screen-pixel–based LOD switching in OSG. (a) 3D view showing the camera frustum, viewport plane and the model bounding sphere; (b) 2D cross-section used for analytical estimation of the model’s projected size, where γ and H represent the vertical field of view and the viewport height in pixels, D denotes the representative model size (we use the bounding-sphere diameter), and d is the distance from the camera to the bounding sphere center.

Comments 2: section 4.3 geometric error and texture error. In regard to the geometric error it is unclear to me to which polygons the Hausdorff distance is applied. Then, I’m wondering if the Chamfer distance could be an alternative. In regard to the texture error I think that the image structural similarity index could provide a much more complete error metric. Image and surface texture have a spatial structure that cannot be captured by just RMS of pixels differences.

Response 2:

We sincerely thank the reviewer for these insightful and constructive suggestions regarding our error metrics. They have significantly improved the rigor and clarity of our evaluation methodology. In our experiments, the Hausdorff distance is computed between the simplified model M₂ and the original, high-resolution model M₁. Both models are converted into dense point clouds (sampled points) for this calculation. The Hausdorff distance is defined as the maximum of the minimum distances between these two point sets. In the Equation (10), h(M₁, M₂) is the one-sided (or directed) Hausdorff distance from M₁ to M₂, and h(M₂), M₁) is the one-sided Hausdorff distance from M₂ to M₁. The final Hausdorff distance is the maximum. We apologize for the omission of “h(M₁, M₂)” in the original manuscript's formula description. This has been corrected in the revised version. While Chamfer distance can serve as an alternative to the Hausdorff distance, we believe the Hausdorff distance is more suitable for our purposes. This is because the Chamfer distance calculates the average mismatch between two point sets, making it less sensitive to outliers. It may tolerate some local large errors as long as the overall average error is small. In contrast, the Hausdorff distance ensures that no point exceeds a certain error threshold, providing a stricter bound on local geometric deviations.

We completely agree with the reviewer that RMSE is insufficient for capturing texture quality. Following your excellent advice, we have revised our texture error metric. It is now defined as 1−SSIM, as detailed in the updated manuscript. This change significantly strengthens our evaluation by ensuring that the reported texture error aligns more closely with human perception of quality loss. We believe this revision has substantially improved the quality and validity of our experimental analysis. Thank you again for this constructive comment.

Author Response File: Author Response.docx

Reviewer 4 Report

Comments and Suggestions for Authors

Thank you for submitting this manuscript. The study focuses on a practically significant and technically challenging issue in digital geology and remote sensing. Overall, the research is clearly structured and demonstrates a certain degree of innovation and application potential, but several aspects, particularly methodological details, experimental validation, and analytical depth, require further improvement.

1. The abstract provides a general overview but lacks quantitative evidence. Please include key performance metrics (e.g., memory reduction ratio, frame rate improvement) to better substantiate the claimed effectiveness.

2. The novelty is not clearly highlighted. Core contributions such as the “vertex sharpness constraint” and “boundary freezing strategy” should be explicitly emphasized to underline the originality of this work.

3. While the introduction outlines the background, the research gap is not convincingly articulated. Since the related work is entirely separated into Section 2, the logical flow between motivation and methodology is weakened. Consider integrating key literature insights into the introduction to strengthen the argument.

4. Section 2 mainly lists previous methods without providing critical evaluation. It would be helpful to clarify why existing approaches (e.g., CityGML or other LOD techniques) fail to handle cliff-like or irregular geological surfaces effectively.

5. Adding a short paragraph or bullet list summarizing the main contributions at the end of the introduction would improve clarity and reader orientation.

6. The LOD generation process (Sections 3.1 & 3.2) is driven by 2D texture space rather than 3D geometric structure. While this benefits texture fidelity, it may not be suitable for highly folded or intersecting surfaces. The limitation is mentioned (Section 5.1) but deserves deeper discussion.

7. The approach relies on several empirical parameters (e.g., Vt=1.0 MB, θ=0.8). These manually tuned settings limit generalization and automation. It is recommended to analyze parameter sensitivity or explore automated selection strategies.

8. Section 5.3 mentions that the LOD construction is time-consuming, but the explanation is superficial. According to the results (Section 4.2), processing Model 2 took approximately 13,428 seconds (~4 hours), which represents a clear bottleneck for an “efficient” method. The authors should analyze the timing breakdown (e.g., QEM computation, texture reconstruction) in more depth.

9. Processing time and CPU utilization (~9%) appear inconsistent. Please clarify the experimental setup and performance measurement methodology to identify possible bottlenecks.

10. Only single-run results are reported. Multiple runs and statistical reporting (mean ± standard deviation) would improve result reliability.

11. The discussion remains somewhat general and should better elaborate on the method’s applicability and limitations for different geological scenarios.

12. The English writing needs polishing, particularly to improve readability, tense consistency, and sentence structure. Professional language editing is advised.

13. The reference list lacks recent studies. Please update with key works from the past three years and ensure consistent citation formatting.

14. The conclusion section summarizes the work well but sounds overly assertive. It should balance claims with limitations and briefly discuss future work, such as automated parameter tuning, parallel optimization, or cloud-based visualization.

Author Response

Comments 1: The abstract provides a general overview but lacks quantitative evidence. Please include key performance metrics (e.g., memory reduction ratio, frame rate improvement) to better substantiate the claimed effectiveness.

Response 1:

We sincerely thank the reviewer for this valuable suggestion. We agree that quantitative evidence is crucial to substantiate the claims of our method's effectiveness. Accordingly, we have revised the abstract to incorporate the key performance metrics as suggested. The updated abstract now explicitly states that:

The models loading time and average memory usage were reduced by more than 90%, while the average display frame rate can reach around 60 FPS.

The revised text is presented in the updated manuscript (please see the highlighted changes in the Abstract section).

Comments 2: The novelty is not clearly highlighted. Core contributions such as the “vertex sharpness constraint” and “boundary freezing strategy” should be explicitly emphasized to underline the originality of this work.

Response 2:

We are grateful to the reviewer for this critical insight. We agree that the novelty of our work needed to be more explicitly articulated. Following the reviewer's suggestion, we have now thoroughly revised the Abstract to highlight our core contributions. In the Abstract and Introduction: The updated abstract now explicitly states that:

The proposed LOD construction method integrates several innovative components: segmentation based on texture coverage, an adaptive LOD tile pyramid using a pseudo-quadtree, and a feature-preserving mesh simplification algorithm that incorporates vertex sharpness constraint and boundary freezing strategy to retain critical geological features.

The revised text is presented in the updated Abstract (please see the highlighted changes in the Abstract section).

 

Comments 3: While the introduction outlines the background, the research gap is not convincingly articulated. Since the related work is entirely separated into Section 2, the logical flow between motivation and methodology is weakened. Consider integrating key literature insights into the introduction to strengthen the argument.

Response 3:

We have revised the introduction to better highlight the research gap by integrating key insights from the related work. Specifically, we emphasize the challenges of current Level of Detail (LOD) technologies in handling irregular outcrop shapes, preserving critical geological features, and efficiently rendering high-resolution data. This revision strengthens the logical flow between the research motivation and methodology, providing a clearer justification for the proposed approach. We believe this will enhance the clarity and coherence of the introduction.  In the revised text, the research gap is now convincingly articulated.

“Their application in digital outcrop modeling still faces three significant challenges: (1) Unsuitable data partitioning for irregular, elongated outcrop shapes; (2) loss of key ge-ological features during model simplification; and (3) inefficient rendering of massive, high-resolution outcrop data.”

 

Comments 4: Section 2 mainly lists previous methods without providing critical evaluation. It would be helpful to clarify why existing approaches (e.g., CityGML or other LOD techniques) fail to handle cliff-like or irregular geological surfaces effectively.

Response 4:

Thank you for your insightful comment. We have revised Section 2 to provide a more critical evaluation of existing methods. Specifically, we now highlight the limitations of approaches like CityGML and other LOD techniques in handling irregular, feature-rich geological surfaces, such as cliffs. These methods, originally designed for urban modeling, often rely on regularized geometric abstractions that struggle to represent the heterogeneous and complex geometries typical of geological outcrops. For example, the discrete LOD levels in CityGML are based on simplified geometric forms like prismatic blocks, which fail to capture the fine-scale geological features, such as fractures and bedding planes, crucial for accurate geological interpretation. Additionally, existing LOD techniques tend to over-smooth or cull essential geological features, resulting in a loss of semantic information. This critical evaluation of the limitations of current methods strengthens the rationale for developing task-oriented, adaptive LOD frameworks tailored to the specific challenges posed by geological outcrop modeling. We believe this revision clarifies the gaps in existing approaches and emphasizes the need for specialized solutions.

 

Comments 5: Adding a short paragraph or bullet list summarizing the main contributions at the end of the introduction would improve clarity and reader orientation.

Response 5:

We thank the reviewer for this valuable suggestion to enhance the clarity and structure of our manuscript. In direct response to this comment, we have thoroughly revised the last paragraph of the Introduction (Section 1) that explicitly summarizes the main contributions of this work in a clear, three-point list:

• A specialized LOD workflow that uses texture-based segmentation and a pseudo-quadtree to generate an adaptive tile pyramid, effectively handling the elongated morphology of outcrops.
• A feature-preserving simplification algorithm that introduces a vertex sharpness constraint and a boundary freezing strategy to retain critical geo-logical features during aggressive mesh reduction.
• A dynamic visualization framework that leverages an LOD index and the OpenSceneGraph (OSG) engine to achieve real-time, view-dependent load-ing and rendering, enabling interactive exploration of massive models on standard hardware.

 

Comments 6: The LOD generation process (Sections 3.1 & 3.2) is driven by 2D texture space rather than 3D geometric structure. While this benefits texture fidelity, it may not be suitable for highly folded or intersecting surfaces. The limitation is mentioned (Section 5.1) but deserves deeper discussion.

Response 6:

We sincerely thank the reviewer for this insightful and valuable comment. We completely agree that a deeper discussion of the limitations of our texture-space approach for complex geometries is crucial. The reviewer has correctly identified a key trade-off in our method. In response to this comment, we have significantly revised and expanded Section 5.1 "Potential Limitations of the Tiling Method" to provide a more thorough and mechanistic analysis of this limitation. The specific revisions can be found in the revised version.

Meanwhile, we have transformed the future work section from a brief mention into a concrete research plan. We now propose a pathway towards a geometry-aware hybrid framework, outlining specific steps: (1) Geometric Feature Analysis using metrics like curvature and normal variation; (2) A Hybrid Quadtree-Octree Partitioning strategy that switches between 2D (for planar regions) and 3D (for complex regions) partitioning based on the geometric analysis; (3) Morphology-Based Classification to automatically select the optimal strategy for different parts of an outcrop model.

Furthermore, to proactively manage reader expectations, we have added a clarifying statement at the beginning of Section 3.2, explicitly noting that our method is optimized for texture fidelity and is best suited for relatively planar or gently curved geometries, with a forward reference to the detailed discussion in Section 5.1.

We believe these revisions have substantially strengthened the paper by providing the deeper discussion the reviewer requested. They demonstrate a clear understanding of the method's boundaries and present a credible vision for future improvements. Thank you again for prompting this important enhancement to our manuscript.

 

Comments 7: The approach relies on several empirical parameters (e.g., Vt=1.0 MB, θ=0.8). These manually tuned settings limit generalization and automation. It is recommended to analyze parameter sensitivity or explore automated selection strategies.

Response 7:

We thank the reviewer for this important observation. We have significantly revised Section 5.2 to address these concerns by: (1) Adding parameter sensitivity analysis: New experiments (Figure 18) show rendering frame rate drops sharply from 59.8 to 17.9 FPS as Vt increases from 1.0 to 4.0 MB, while θ=0.8 is shown to be optimal based on error curves in Figure 16; and (1) We now outline concrete approaches for dynamic parameter selection: Vt adapting to hardware/network conditions, and θ varying locally based on geometric/textural features and semantic importance. Please see the revised Section 5.2 for specific modifications.

 

Comments 8: Section 5.3 mentions that the LOD construction is time-consuming, but the explanation is superficial. According to the results (Section 4.2), processing Model 2 took approximately 13,428 seconds (~4 hours), which represents a clear bottleneck for an “efficient” method. The authors should analyze the timing breakdown (e.g., QEM computation, texture reconstruction) in more depth.

Response 8:

We sincerely thank the reviewer for this critical observation regarding the computational efficiency of our LOD construction process. We agree that the significant processing time reported in our initial submission represents a bottleneck that warrants deeper investigation.

In direct response to this comment, we have substantially revised Section 5.3, now titled "Analysis of Algorithmic Performance Bottlenecks and Optimization Directions" to provide a detailed and transparent analysis. The key enhancements include:

(1) Detailed Timing Breakdown: We have performed comprehensive profiling of the LOD construction pipeline. The results are presented in the new Figure 19, which clearly delineates the time distribution across four key stages: Data I/O & Initialization, Texture Processing, Mesh Simplification, and Segmentation & Tiling.

(2) Explicit Bottleneck Identification: Based on empirical data, we clearly identify and analyze the primary bottlenecks: (1) Data I/O and Initialization, accounting for 64% and 68% of total time in Model 1 and Model 2 respectively, revealing critical inefficiencies in handling massive datasets in our current implementation; and (2) Texture Processing, which demonstrates an order of magnitude higher time consumption than mesh simplification for high-resolution texture models.

(3) Concrete Optimization Framework: We have proposed a specific optimization strategy that moves beyond general parallelization concepts to include: 1) Implementation of pipelining to overlap data I/O, geometric processing, and texture operationsWe believe this significantly strengthens the paper's critical analysis and future outlook; 2) GPU-based parallelization of texture operations (partitioning, downsampling, and remapping); 3)Restructuring the QEM algorithm for parallel vertex error calculations and edge collapse evaluations on the GPU.

 

Comments 9: Processing time and CPU utilization (~9%) appear inconsistent. Please clarify the experimental setup and performance measurement methodology to identify possible bottlenecks.

Response 9:

The processing time was measured by recording the total time of the LOD construction workflow. CPU utilization and memory usage were monitored at the system level using the Windows Resource Monitor, reporting the average utilization across all logical cores during the entire process. The frame rate was observed using methods provided by the OSG engine.

The apparent discrepancy between the total processing time (~4 hours for Model 2) and the reported average CPU utilization (~12%) can be explained by two primary factors. First, our core simplification algorithm is predominantly single-threaded. On the test platform, which uses a 16-thread CPU, a perfectly optimized single-threaded task would yield a maximum average utilization of approximately 6.25% (1/16), which aligns closely with the observed value. The remaining utilization comes from parallelized ancillary tasks (such as texture handling, file I/O) and system processes. Second, significant portions of the pipeline, especially data I/O (loading the initial large mesh and writing the generated LOD tiles), are I/O-bound operations. During these phases, the CPU often remains idle while waiting for data transfers, which further lowers the overall utilization.

 

This indicates that the bottleneck is not related to computational resources but to the inherent sequential nature of the core algorithm and the limitations imposed by disk I/O speeds. This issue is discussed in detail in Section 5.3, where we provide further insights into the I/O-related bottlenecks and potential optimizations.

 

Comments 10: Only single-run results are reported. Multiple runs and statistical reporting (mean ± standard deviation) would improve result reliability.

Response 10:

We thank the reviewer for this crucial suggestion to enhance the statistical rigor of our results. We completely agree that multiple runs are essential for demonstrating reliability.

In direct response to this comment, we have repeated all key experiments five times and have thoroughly revised the Results section (Section 4). All performance metrics, including loading time, memory usage, and frame rate, are now reported as the mean ± standard deviation. The results consistently show very low standard deviations, which confirms the high stability and repeatability of our method.

 

Comments 11: The discussion remains somewhat general and should better elaborate on the method’s applicability and limitations for different geological scenarios.

Response 11:

We agree with the reviewer that delineating the method's boundaries is crucial. In response, we have significantly expanded the discussion in Section 5 to include a dedicated analysis of its applicability and limitations. We have systematically addressed the following aspects in the revised discussion:

5.1 Potential Limitations of the Tiling Method: We have discussed fundamental limitations of our texture-space partitioning approach when handling complex geological geometries, including detailed discussions on suboptimal tile boundaries and inefficient culling in folded structures.

5.2 Empirical and Generalizability Concerns in Algorithm Parameters: We now provide empirical evidence and sensitivity analysis for our parameter selections, along with proposing a data-driven framework for automated parameter optimization in future work.

5.3 Analysis of Algorithmic Performance Bottlenecks and Optimization Directions: A comprehensive analysis of algorithmic performance identifies specific computational bottlenecks and presents optimization strategies, including parallelization and pipelining approaches that could reduce processing time from hours to minutes.

 

Comments 12: The English writing needs polishing, particularly to improve readability, tense consistency, and sentence structure. Professional language editing is advised.

Response 12:

We thank the reviewer for highlighting the need for language improvement. We agree that clear and polished writing is essential for effective scientific communication. In response, the entire manuscript has been professionally edited by Yaxiong Shao (Northern Illinois University, USA), addressing the language concerns.

 

Comments 13: The reference list lacks recent studies. Please update with key works from the past three years and ensure consistent citation formatting.

Response 13:

In response to the reviewer's comment, we have revised the reference list by incorporating several seminal works from the past three years and have performed a comprehensive check to guarantee consistent formatting for all citations.

 

Comments 14: The conclusion section summarizes the work well but sounds overly assertive. It should balance claims with limitations and briefly discuss future work, such as automated parameter tuning, parallel optimization, or cloud-based visualization.

Response 14:

We thank the reviewer for this constructive suggestion. We have thoroughly revised the Conclusion section to temper the assertive tone and to incorporate a balanced discussion of the study's limitations and future work. The revised conclusion now explicitly acknowledges the limitations of our approach, particularly its reliance on empirically tuned parameters and the computational intensity of the preprocessing stage. Furthermore, we have integrated a concise outlook on specific future research directions, including automated parameter tuning, parallel optimization of the core algorithms, and enhanced handling of geometrically complex scenarios, as recommended. The revised Conclusion has been highlighted in the manuscript for the reviewer's convenience.

Author Response File: Author Response.docx

Round 2

Reviewer 4 Report

Comments and Suggestions for Authors

The authors have satisfactorily addressed all concerns I raised, and the manuscript quality is significantly improved. I recommend accepting it in its present form.