Multi-Technique 3D Modelling of Narrow Gorges to Assess Stability: Case Study of Caminito Del Rey (Spain)

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Dear authors,

I service as the reviewer of your manuscript “Multi-technique 3D modelling of narrow gorges to assess stability. Case study of Caminito del Rey (Spain)”. It proposes a methodological approach to integrate digital photogrammetry and laser data, leveraging their respective strengths to generate reliable products for landslide risk assessment. Authors claim that the method shows promise as a data-driven tool for rockfall susceptibility. However, due to some limitations, I think the present study should be revised before it can be considered in the journal Remote Sensing. Below you can find my concerns:

- There is no quantitative results in the Abstract. Authors present many backgrounds and methods, but with insufficient results. In addition, the major novelty of this study should be mentioned in this section.

- The paper structure is a big issue. You should add a “study area” or “materials” section. Then you could put “Caminito del Rey” (including Figure 1) into that section not in Introduction.

- Section 2.1 should be more specific. In my opinion, you mix too much information in this part, which should be separated. First, you may mention the overall framework of your analysis. Then you must list some details of the used methods, especially important equations. It would be better to add a third-level heading for this part.

- Terminology is very important. You use both landslides and rockfalls in the text, which makes me confused. You may mention somewhere (especially at beginning) the proposed method is useful for the landslides, but a rockfall is used as a case study. This would be more clear. All the “landslides” in the text should be discussed in depth.

- Figure 4. Some details must be added. For example, what do the different colors represent in (d)? In (a) and (f), the meanings of the lines (or points?) is not clear in the figure.

- Figure 5 is the most important output for your study! However, it is not clear, and some important information misses. For example, the scale. The subfigure (f) doesn’t provide useful information. This must be improved.

- Discussion section should be apart from Conclusion. In Discussion, you can list the limitations or uncertainties of the present method, and also compare with previous studies.

Author Response

Reviewer #1

Dear authors,

I service as the reviewer of your manuscript “Multi-technique 3D modelling of narrow gorges to assess stability. Case study of Caminito del Rey (Spain)”. It proposes a methodological approach to integrate digital photogrammetry and laser data, leveraging their respective strengths to generate reliable products for landslide risk assessment. Authors claim that the method shows promise as a data-driven tool for rockfall susceptibility. However, due to some limitations, I think the present study should be revised before it can be considered in the journal Remote Sensing. Below you can find my concerns:

- There is no quantitative results in the Abstract. Authors present many backgrounds and methods, but with insufficient results. In addition, the major novelty of this study should be mentioned in this section.

Authors: The abstract has been improved following your suggestions. We have integrated the main results into the text, highlighting the comprehensive products and the high-resolutions achieved. The novelty of the multi-sensor cohesive integration framework has been explicitly emphasized.

Reviewer #1:

- The paper structure is a big issue. You should add a “study area” or “materials” section. Then you could put “Caminito del Rey” (including Figure 1) into that section not in Introduction.

Authors: We have included a specific “Case Study” section (Section 1.1) following your suggestion, integrating the content (including Figure 1). We have also added a specific discussion on the limitations imposed by the study area.

- Section 2.1 should be more specific. In my opinion, you mix too much information in this part, which should be separated. First, you may mention the overall framework of your analysis. Then you must list some details of the used methods, especially important equations. It would be better to add a third-level heading for this part.

Authors: We have reorganized the structure of the paper to address this concern. The Methodology is now described under its own header, and the Application has been integrated into the Results section. Both sections utilize second-level headings to clarify the overall framework, separate concepts, and detail the specific methods used, as suggested. Thank you for your suggestion.

- Terminology is very important. You use both landslides and rockfalls in the text, which makes me confused. You may mention somewhere (especially at beginning) the proposed method is useful for the landslides, but a rockfall is used as a case study. This would be more clear. All the “landslides” in the text should be discussed in depth.

Authors: Rockfalls are a common type of landslide [1]. However, to avoid confusion, and since our approach is primarily focused on rockfall assessment (which is common in narrow gorges and canyons), we have changed all instances of the term 'landslide' to 'rockfall' throughout the manuscript.

- Figure 4. Some details must be added. For example, what do the different colors represent in (d)? In (a) and (f), the meanings of the lines (or points?) is not clear in the figure.

Authors: We have improved Figure 4 by adding more specific details to the caption regarding the meaning of colors, lines, and symbols for better clarity. Thank you.

- Figure 5 is the most important output for your study! However, it is not clear, and some important information misses. For example, the scale. The subfigure (f) doesn’t provide useful information. This must be improved.

Authors: We have improved Figure 5 to enhance its clarity and ensure scale information is apparent where necessary. We have improved subfigure (f) showing the point cloud visualization tool that are integrated in the geoportal. Subfigure (f) remains as it provides a crucial example of the final product: the geoportal (WebGIS platform). This geoportal facilitates visualization of the point cloud and digitalized features (like the walkway), demonstrating the practical utility of our products for geologists and engineers in analyzing potential rockfall events.

- Discussion section should be apart from Conclusion. In Discussion, you can list the limitations or uncertainties of the present method, and also compare with previous studies.

Authors: We have created a dedicated Discussion section (Section 4) separate from the Conclusions (Section 5), as suggested. We have improved the Discussion to include sub-sections detailing the limitations, uncertainties, and comparative analysis of our approach.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

The submitted paper presents an innovative multi-technology data fusion methodology that integrates four core technologies: UAV-LiDAR, TLS, MMS, and SP. The method is specifically designed to address 3D modeling challenges in highly occluded, complex terrains, such as narrow canyons and vertical rock walls. It effectively fills the application gap left by single technologies or simple combinations, achieving seamless “no blind spot” 3D recording. The generated point cloud achieves an average density of 1–10 cm, with over 20 billion points—overcoming traditional limitations in both data density and coverage. The submitted paper highlights the “collaborative operation of large and small UAVs” and introduces “extrinsic parameter calibration of 360° cameras.” By pre-calibrating fixed distances between camera sensors and using them as “scale constraints,” the method significantly reduces reliance on GCPs. Overall, this study represents an essential advancement in integrating airborne and ground-based multi-sensing technologies for high-quality 3D modeling. However, the current version still requires substantial improvement in both structure and technical detail, as outlined below:

Main points:

1. The introduction section should elaborate on the unique challenges of rockfall monitoring in narrow canyon terrain—such as rock wall occlusion, limited space, and weak GNSS signals—and clearly state the specific limitations of existing single technologies in these environments. For example, TLS often lacks data coverage in canyon areas obstructed by vegetation, while UAVs may struggle to maneuver in confined spaces. Providing such examples would better highlight the necessity and innovation of the proposed multi-technology fusion approach.
2. It is recommended to incorporate recent research (2020–2025) on multi-technology fusion in narrow canyons or similarly complex terrains. The discussion should analyze fusion logic (e.g., data complementarity, accuracy matching), current limitations (e.g., redundant data handling, differing coordinate systems), and how this study advances beyond prior work. Avoid merely listing references—focus instead on a critical comparison, particularly regarding the “360° camera parameter calibration” technique used in this paper.
3. The co-authors should include more details on the design of flight routes and scanning stations, considering the specific geomorphological features of canyon terrain, and for example, increasing the number of stations in narrow zones or adjusting scanning angles for vertical rock faces. Data coverage should be verified (e.g., by identifying occlusion zones via simulations or pre-scans). A clear workflow diagram showing data acquisition steps, efficiency metrics, and time requirements for each component is highly recommended to improve transparency and reproducibility.
4. Although the paper mentions establishing a reference system via a GNSS network, it lacks details on how data from UAV-LiDAR, TLS, and MMS are transformed into a unified coordinate system. The authors should describe whether common control points are used and which error-correction methods are applied. It is suggested to include mathematical models for coordinate transformations (e.g., 7-parameter or 3-parameter transformations) and to explain the accuracy verification process (e.g., deviations computed from checkpoints). This will ensure clarity and control in the data fusion process.
5. The discussion section should expand on how the resulting 3D model can be practically applied to improve safety management on narrow mountain roads. The authors should also discuss the adaptability of this methodology to other similar scenarios—such as canyon roads or slopes in water conservancy projects—and outline necessary methodological adjustments. For instance, UAV flight paths in road canyons may require optimization to minimize traffic interference, while water conservancy areas must account for LiDAR reflections from water surfaces. The effects of noise in data and potential denoising algorithms should also be addressed to enhance the practical applicability and generalizability of the research.
6. In Figure 5, the co-authors demonstrate the 3D model results. It is recommended to include an additional panoramic view of the study area, indicating the flight paths and labeling potential or historical rockfall zones. This addition would significantly strengthen the paper’s applied value by demonstrating how the model supports hazard identification rather than simply illustrating data acquisition.

Minor points:

      1. The references section should strictly follow the MDPI formatting guidelines.

      2. Abbreviations should be standardized; only technological terms should be abbreviated in keywords (application scenarios should be written in full).

      3. Figure legends and annotations should be improved for clarity.

      4. For Figure 1, a simple DEM background is recommended to replace the current overly complex textual base.

 

Author Response

Reviewer #2:

The submitted paper presents an innovative multi-technology data fusion methodology that integrates four core technologies: UAV-LiDAR, TLS, MMS, and SP. The method is specifically designed to address 3D modeling challenges in highly occluded, complex terrains, such as narrow canyons and vertical rock walls. It effectively fills the application gap left by single technologies or simple combinations, achieving seamless “no blind spot” 3D recording. The generated point cloud achieves an average density of 1–10 cm, with over 20 billion points—overcoming traditional limitations in both data density and coverage. The submitted paper highlights the “collaborative operation of large and small UAVs” and introduces “extrinsic parameter calibration of 360° cameras.” By pre-calibrating fixed distances between camera sensors and using them as “scale constraints,” the method significantly reduces reliance on GCPs. Overall, this study represents an essential advancement in integrating airborne and ground-based multi-sensing technologies for high-quality 3D modeling. However, the current version still requires substantial improvement in both structure and technical detail, as outlined below:

Main points:

1. The introduction section should elaborate on the unique challenges of rockfall monitoring in narrow canyon terrain—such as rock wall occlusion, limited space, and weak GNSS signals—and clearly state the specific limitations of existing single technologies in these environments. For example, TLS often lacks data coverage in canyon areas obstructed by vegetation, while UAVs may struggle to maneuver in confined spaces. Providing such examples would better highlight the necessity and innovation of the proposed multi-technology fusion approach.

Authors: We agree with this suggestion. We have emphasized the unique challenges of rockfall monitoring in narrow gorges by including a new paragraph in the Introduction, prior to describing the study area, detailing the specific limitations of single technologies in these environments (occlusion, limited space, weak GNSS signals).

Reviewer #2:

1. It is recommended to incorporate recent research (2020–2025) on multi-technology fusion in narrow canyons or similarly complex terrains. The discussion should analyze fusion logic (e.g., data complementarity, accuracy matching), current limitations (e.g., redundant data handling, differing coordinate systems), and how this study advances beyond prior work. Avoid merely listing references—focus instead on a critical comparison, particularly regarding the “360° camera parameter calibration” technique used in this paper.

Authors: We have included recent research (2020–2025) related to multi-technology fusion in complex scenes. We confirm that the bibliographic search did not yield recent work that addresses the cohesive fusion of these specific technologies (aerial and terrestrial, static and mobile, conventional and spherical photogrammetry) in a single narrow gorge scenario, which underscores the novelty and the gap covered by our methodological framework. Thank you.

Reviewer #2:

1. The co-authors should include more details on the design of flight routes and scanning stations, considering the specific geomorphological features of canyon terrain, and for example, increasing the number of stations in narrow zones or adjusting scanning angles for vertical rock faces. Data coverage should be verified (e.g., by identifying occlusion zones via simulations or pre-scans). A clear workflow diagram showing data acquisition steps, efficiency metrics, and time requirements for each component is highly recommended to improve transparency and reproducibility.

Authors: We appreciate this suggestion. Flight routes were planned and simulated using PNOA data and an in-house application [31], with the resulting plans incorporated into the operational software. The positions of scanning stations (TLS and MMS) and the 360-degree camera were defined in the field to ensure high overlap and coverage in specific zones, such as narrow sections and vertical faces. This planning detail has been added to the Data Acquisition (Section 3.1) section for clarity and transparency.

Reviewer #2:

1. Although the paper mentions establishing a reference system via a GNSS network, it lacks details on how data from UAV-LiDAR, TLS, and MMS are transformed into a unified coordinate system. The authors should describe whether common control points are used and which error-correction methods are applied. It is suggested to include mathematical models for coordinate transformations (e.g., 7-parameter or 3-parameter transformations) and to explain the accuracy verification process (e.g., deviations computed from checkpoints). This will ensure clarity and control in the data fusion process.

Authors: We have substantially improved Section 2.1 to clarify the coordinate system definition and fusion process. We confirm that we have detailed the use of second-order control points and the rigid 3D transformation employed to unify the datasets. Furthermore, the Data Processing (Section 3.2) section now explicitly describes the accuracy verification process from the independent Checkpoints (CPs), thus ensuring clarity and control over the data fusion process.

Reviewer #2:

1. The discussion section should expand on how the resulting 3D model can be practically applied to improve safety management on narrow mountain roads. The authors should also discuss the adaptability of this methodology to other similar scenarios—such as canyon roads or slopes in water conservancy projects—and outline necessary methodological adjustments. For instance, UAV flight paths in road canyons may require optimization to minimize traffic interference, while water conservancy areas must account for LiDAR reflections from water surfaces. The effects of noise in data and potential denoising algorithms should also be addressed to enhance the practical applicability and generalizability of the research.

Authors: We have expanded the discussion on the practical application of the resulting 3D model in Section 4.3 (Generalizability and practical application), specifically discussing its adaptability to other scenarios such as canyon roads and slopes in water conservancy projects. We also discuss noise management and the need for future spatially-explicit uncertainty analysis to enhance generalizability. Thank you for your suggestion.

Reviewer #2:

1. In Figure 5, the co-authors demonstrate the 3D model results. It is recommended to include an additional panoramic view of the study area, indicating the flight paths and labeling potential or historical rockfall zones. This addition would significantly strengthen the paper’s applied value by demonstrating how the model supports hazard identification rather than simply illustrating data acquisition.

Authors: We have improved Figure 5 by including a reference/labeling an area where historical rockfall events were reported. This strengthens the paper’s applied value by demonstrating how the model supports hazard identification.

Reviewer #2:

Minor points:

1. The references section should strictly follow the MDPI formatting guidelines.
2. Abbreviations should be standardized; only technological terms should be abbreviated in keywords (application scenarios should be written in full).
3. Figure legends and annotations should be improved for clarity.
4. For Figure 1, a simple DEM background is recommended to replace the current overly complex textual base.

Authors: Done. Thank you.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

All comments are presented in the text document

Comments for author File: Comments.pdf

Author Response

Reviewer #3:

Authors: Dear reviewer, we have included your suggestions in the manuscript. We have also re-organized the structure for clarity using multiple sections and sub-sections. The figures have also been modified following your indications. Thank you.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

I think all my concerns have been addressed well. I agree to accept it.

Reviewer 3 Report

Comments and Suggestions for Authors

No comments