Performance Assessment of Portable SLAM-Based Systems for 3D Documentation of Historic Built Heritage
Round 1
Reviewer 1 Report
Comments and Suggestions for Authors
The article shows two interesting applications of SLAM technology to two religious buildings with the same architectural characteristics. The data acquisition and processing work appears rigorous and well done, with a good level of detail in the explanation of the individual activities carried out. The comparison of the data returned, which is essential for the correct formulation of the proposed results, also appears to be well done and focused on the most important points in justifying the results. It now seems interesting to apply it in terms of multidisciplinarity, for example by linking it to a critical reading of the stratigraphy of the monument in order to effectively understand its potential in this field as well.
Author Response
We thank the reviewer for the positive evaluation. We agree that a stratigraphic reading of masonry structures can contribute to a more detailed understanding of construction phases and structural behaviour, potentially supporting seismic vulnerability assessment. However, the present study focuses on preliminary assessment procedures, for which such detailed analyses are not strictly required. The proposed SLAM-based 3D documentation nonetheless provides a solid geometric basis that could support more detailed multidisciplinary and structural analyses, including stratigraphic investigations, in future research.
Reviewer 2 Report
Comments and Suggestions for Authors
This paper investigats the utility of three SLAM systems for the rapid geometric documentation of historic built heritage, specifically focusing on two churches in the Lunigiana region. The goal is to assess the potential of these tools to support preliminary and simplified seismic vulnerability assessments. The results demonstrate the capability of SLAM systems to efficiently capture primary geometries. However, the connection between the achieved metric accuracy and the specific requirements of the seismic vulnerability assessment method needs to be strengthened. Some suggestions are provided for the authors' consideration.
- The paper claims to support seismic vulnerability assessment, but it does not explicitly state how the achieved metric accuracy satisfies the input demands of any specific simplified structural analysis method.
- The descriptions of the two churches are currently focused on their historical background. For a study related to seismic vulnerability, it would be beneficial to enrich the structural characterization, including the type, condition of masonry, structural system layouts.
- Since the paper positions itself within the seismic field, the Introduction would benefit from references to relevant structural engineering literature. As the core contribution concerns geometric data acquisition, it would be logical to highlight that a structure’s seismic performance is fundamentally governed by geometric characteristics of members and connection details, which control both its elastic behaviour, seismic and post-earthquake behaviour. The authors could include some recent seismic studies, such as “Fuse replacement implementation by shaking table tests on hybrid moment-resisting frame” “Study of the strength and behaviour of single-coped beam under combined bending, shear and axial loads”, “Closed-form design solutions for parallelogram hollow structural sections under bending scenarios ”, etc.
- Historic masonry structures are irregular. SLAM point clouds often have significant noise. A qualitative or quantitative assessment of the data noise/scatter for each SLAM system is suggested. It would allow readers to better assess the reliability of SLAM-derived geometries for engineering applications
Author Response
This paper investigats the utility of three SLAM systems for the rapid geometric documentation of historic built heritage, specifically focusing on two churches in the Lunigiana region. The goal is to assess the potential of these tools to support preliminary and simplified seismic vulnerability assessments. The results demonstrate the capability of SLAM systems to efficiently capture primary geometries. However, the connection between the achieved metric accuracy and the specific requirements of the seismic vulnerability assessment method needs to be strengthened. Some suggestions are provided for the authors' consideration.
- The paper claims to support seismic vulnerability assessment, but it does not explicitly state how the achieved metric accuracy satisfies the input demands of any specific simplified structural analysis method.
We thank the reviewer for this comment. We would clarify that the study is part of a broader research project aimed at supporting seismic vulnerability assessment of historic buildings. However, in this specific manuscript, seismic vulnerability aspects are addressed only marginally and mainly as an application context. This aspect has been explicitly discussed at the end of Section 4.2 and Section 4.3.2, and further stressed in the Conclusions.
The descriptions of the two churches are currently focused on their historical background. For a study related to seismic vulnerability, it would be beneficial to enrich the structural characterization, including the type, condition of masonry, structural system layouts.
We have expanded the description of the first of the two churches chosen as case studies, paying greater attention to the structural aspects of the buildings. As far as Vallecchia church is concerned, we believe that the structural characteristics have already been detailed.
2. Since the paper positions itself within the seismic field, the Introduction would benefit from references to relevant structural engineering literature. As the core contribution concerns geometric data acquisition, it would be logical to highlight that a structure’s seismic performance is fundamentally governed by geometric characteristics of members and connection details, which control both its elastic behaviour, seismic and post-earthquake behaviour. The authors could include some recent seismic studies, such as “Fuse replacement implementation by shaking table tests on hybrid moment-resisting frame” “Study of the strength and behaviour of single-coped beam under combined bending, shear and axial loads”, “Closed-form design solutions for parallelogram hollow structural sections under bending scenarios ”, etc.
The Introduction has been completely revised and reorganized, separating it from the literature review. The State of the Art section has been expanded with additional references, including one of those suggested by the reviewer. The other proposed references, focused on laboratory testing of metal structures, were considered not directly relevant to the scope of this work.
3. Historic masonry structures are irregular. SLAM point clouds often have significant noise. A qualitative or quantitative assessment of the data noise/scatter for each SLAM system is suggested. It would allow readers to better assess the reliability of SLAM-derived geometries for engineering applications.
We have added a qualitative discussion in Section 4.3.2 highlighting the impact of noise on model readability, which in several cases affects interpretation more than resolution itself. A more detailed quantitative analysis of noise is currently under development and will be addressed in future work.
Author Response File: 
Author Response.pdf
Reviewer 3 Report
Comments and Suggestions for Authors
Please see the attachment.
Comments for author File: Comments.pdf
Author Response
The research investigates the potential and limitations of SLAM technology for the precise reconstruction of cultural heritage, using two churches as case studies. The results highlight the measured performance of the different instruments considered, with the best results obtained using the VLX 3 and the lowest performance observed with the Lixel L2 Pro. In light of recent technological advancements in SLAM systems, the study provides a relevant and timely contribution to the field. The methodological approach is sound and is presented in a comprehensive manner throughout the manuscript.
However, several important issues should be addressed. First, the title of the paper does not accurately reflect the scope of the research. The methodology primarily focuses on experimental testing of different SLAM-based instruments applied to cultural heritage objects. These experiments include detailed analyses of georeferencing accuracy, 3D model accuracy, model density and resolution, point cloud density and resolution, and 3D model completeness. While these analyses are appropriate for evaluating the performance of SLAM technologies in the precise reconstruction of real-world structures, the manuscript does not include any analysis related to the seismic vulnerability assessment of historic built heritage, which is explicitly referenced in the paper title. Consequently, the title is misleading and should be revised to more accurately represent the actual content and focus of the study.
We acknowledge the reviewer’s comment and clarify that the study is part of a broader research project aimed at supporting seismic vulnerability assessment of historic buildings. However, in this specific manuscript, seismic vulnerability aspects are addressed only marginally and mainly as an application context. To accurately reflect the actual content and avoid any misleading interpretation, the title has been revised to:
“Performance Assessment of Portable SLAM-Based Systems for 3D Documentation of Historic Built Heritage”
The second concern relates to the academic organization of the manuscript. Standard scientific writing practice typically involves first establishing the broader research context and problem through a literature-based discussion, followed by a clear definition of the research objectives and the applied methodology. This issue is particularly evident in the Abstract and Introduction sections. The abstract begins directly with the subject of the study without sufficiently positioning it within a broader research context. Similarly, the introduction interweaves the state of the art with the specific objectives of the study, which affects clarity and readability. Given that the introduction is currently lengthy and difficult to follow, I recommend introducing a dedicated Section 2 (Literature Review) and reorganizing much of the existing content accordingly. The Introduction should then focus on defining the broader research context, the identified research gap, the objectives of the study, and its methodological approach and contribution to the field.
We have fully reorganized the manuscript structure. The Introduction has been revised to focus on the research context, objectives, and contribution, while a dedicated Section 2 (“State of the Art and Literature Review”) has been introduced to improve clarity and readability.
Although the literature review includes a substantial number of references, it would benefit from the inclusion of additional studies that are more directly related to the reconstruction of complex geometric structures in church heritage. Relevant examples include:
- Pérez-García, J.L., Gómez-López, J.M., Mozas-Calvache, A.T. and DelgadoGarcía, J., 2024. Analysis of the photogrammetric use of 360-degree cameras in complex heritage-related scenes: case of the Necropolis of Qubbet el-Hawa (Aswan Egypt). Sensors, 24(7), p.2268.
- Đurić, I., Obradović, R., Vasiljević, I., Ralević, N., & Stojaković, V. (2021). Two-dimensional shape analysis of complex geometry based on photogrammetric models of iconostases. Applied Sciences, 11(15), 7042.
We have added the suggested references and integrated them into the revised literature review. These works are now explicitly discussed in relation to complex heritage documentation and advanced photogrammetric approaches.
Finally, the conclusion section should be revised to more explicitly address the limitations of the study, the encountered challenges, and potential directions for future research.
The Conclusions section has been expanded.
Overall, the study presents thorough experimental work and provides concrete results on the performance of different SLAM instruments applied to church heritage, representing an important contribution to the field. The paper shows clear potential, nevertheless, addressing the issues outlined above would significantly enhance its academic value and publishability.
Author Response File: Author Response.pdf
Reviewer 4 Report
Comments and Suggestions for Authors
Please find attached the manuscript with comments.
Comments for author File: Comments.pdf
Comments on the Quality of English Language
A native English speaker should read the text.
Author Response
Thank you very much for taking the time to review this manuscript. Please find the detailed responses below and the corresponding revisions highlighted in the re-submitted files.
- We have carefully checked the full text to ensure that acronyms are spelled out when they first appear.
- We insert the reference to the interesting paper from Hamzic et al.
[28] - The Introduction has been completely revised and reorganized, separating it from the literature review.
[122] - corrected
[193] - We have expanded the description of the first of the two churches chosen as case studies, paying greater attention to the structural aspects of the buildings, and added a photo framing the bell tower.
- We improved section 3.2.1 The off-line processing according to the Reviewer suggestions; in particular we added Table 2 to summarize the adopted software key features. We clarify different options available for data processing. Paths lengths were already listed in Table 4 - we insert a reference in the text.
[333] - Rephrased: However, the software has limited technical documentation and offers few options to adjust or optimize parameters, which restricts control over the results.
[379] - The 3D model shows some relevant differences in point density, as better described in section 4.3. In the GOPOST software for X70GO, the workflow consists of calculating the cloud from SLAM data, then optimizing it with a noise-reduction filter. In this step, there is an option for “densification,” an oversampling method that depends on the resolution of the original cloud and produces a model three times denser. In the software version adopted for the data processing presented here (November/December 2024), this option was not optional. IVION enable the automatic blurring of people’s faces and plates on the texture, a key aspect when privacy protection is crucial, while GOPOST and LIXEL STUDIO includes automatic removal of people and moving objects.
[424] - At the beginning of section 4.1. Analysis of Georeferencing Accuracy we clarify the role of Figures 5 and 6 and Table 4
[490] - Regarding the residuals in the Sorano church apse, they may be related to issues with total station measurements, as the lines of sight were very foreshortened for those points.
[Table 6 and 504] - When the same route was processed with both targets and GNSS, the latter solution produced slightly worse results, demonstrating the better accuracy provided by georeferencing supported by a topographic survey compared to the GNSS option.
[536] - rephrased: Point cloud density is a critical factor in determining the ability to read fine details in the point cloud: too low a density results in striping artifacts and small features may be missed.
[571] - the reference to the figures has been added
[581] - rephrased and insert more detailed references to the figures: Subsequently, the left inner wall of the church was also analyzed for both datasets; the area was scanned multiple times with both instruments, following homogeneous acquisition paths, obtaining a similar total number of points in the X70GO and VLX 3 datasets
[Figure 11] - improved the caption and insert more detailed references to the Figures in the text: Density evaluation results: Sorano Church - internal left wall: (a) X70GO data, acquired in the path S04/24; (b) VLX 3 data, acquired in the path V01/25. The graphs (c) and (d) show the number of neighborhood points within a 5 cm radius sphere: the less dispersed the Gaussian distribution, the more homogeneous the point density.
[647] - Elaborated the reasoning about VLX3 data higher consistency: Despite having a lower point density than the L2PRO dataset, the VLX3 acquisition exhibits superior visual interpretability of architectural details. This improved readability is attributed to higher local geometric consistency and reduced spatial noise, which enhance the definition of edges, surface continuity, and decorative elements. Conversely, the higher point density of the L2PRO data tends to amplify residual SLAM misalignments and measurement noise, resulting in thicker surfaces and less distinct fine details. These results indicate that point density alone is not a reliable indicator of data quality for detailed architectural analysis.
[714] We avoid reference to NURBS by rewriting: The reconstructed surfaces, generated by profile extrusion, have then been transformed into point clouds, with a resolution comparable to that of the models to be tested.
[Figure 18] and [Table 7] - We insert comments about the results and the table in the text.
[742] - We explained why there is a need for two reference models to test completeness: Since the model deformations exceed the threshold used to assess data completeness, comparison with the reference model constructed from X70 and VLX3 data yields insignificant results for completeness assessment. In this case, as we only wanted to consider the completeness, not the accuracy, of the model, the reference surfaces were adapted to the Lixel data, and the comparison was then carried out as described for the other datasets.
[Table 8] - We insert comments about the results and the table in the text.
Author Response File: Author Response.pdf
Round 2
Reviewer 3 Report
Comments and Suggestions for Authors
The manuscript has been significantly improved compared to the previous version. I would like to thank the authors for addressing my suggestions. I have no further comments and recommend the manuscript for further processing.
Reviewer 4 Report
Comments and Suggestions for Authors
The authors have corrected all that was needed. The paper can be accepted in this form.
