3D Point Cloud to BIM: Semi-Automated Framework to Define IFC Alignment Entities from MLS-Acquired LiDAR Data of Highway Roads

Round 1

Reviewer 1 Report

The article presents a semi-automated method for detection and classification of road markings, that takes as input a raw point cloud from a Mobile Mapping System, and outputs an IFC-compliant file. This method models the alignment and the centreline of each road lane in a highway road. The point cloud processing methodology is validated for two steps: road marking processing and road line extraction. The authors designed a UML diagram for the definition of the alignment entity from the point cloud data. The article contains a very detailed description of the technology of task execution. The research methodology covers the majority of the study. Less emphasis has been put on comparing the obtained results to the work of other authors, including publications on the effectiveness of creating other layers for IFC.

Author Response

Reviewer 1

The article presents a semi-automated method for detection and classification of road markings, that takes as input a raw point cloud from a Mobile Mapping System, and outputs an IFC-compliant file. This method models the alignment and the centreline of each road lane in a highway road. The point cloud processing methodology is validated for two steps: road marking processing and road line extraction. The authors designed a UML diagram for the definition of the alignment entity from the point cloud data. The article contains a very detailed description of the technology of task execution. The research methodology covers the majority of the study.

> Authors want to thank the reviewer for his/her comments, and the effort and time employed reviewing this manuscript.

Less emphasis has been put on comparing the obtained results to the work of other authors, including publications on the effectiveness of creating other layers for IFC.

> Thanks for the comment. Regarding point cloud processing, as it is stated in the Introduction, “Note that this road marking processing method does not aim to be a contribution by itself, but it is essential for the whole workflow and will be validated to prove that it has state-of-the-art performance.” Authors believe that a comparison with other road marking processing works is not necessary in the context of this publication, as the target of the process differs from the state-of-the-art works (which generally aim for a more complete and broad definition of more types of road markings). Therefore, by showing the results of the validation for our case study data, the objective is to remark the feasibility of the method for extracting valuable information to define the IFC file.

The discussion about the inclusion of other IFC layers is interesting.  In the Discussion Section, it can be read that “[…] different infrastructures could be interconnected (railways, bridges,etc.) as a result of the harmonization process that is at the core of IFC 4.3. As the standardization process evolves, different civil engineering software tools are expected to be able to work with IFC files, hence the interest and potential of point cloud processing tools that allow the generation of this type of information models.” At the present time, IFC standards for infrastructure (for the context of this paper, IFC Road) have been recently accepted as Candidate Standard by BuildingSmart International.  That means that they are still under development, so work related to creating other IFC layers in road infrastructure context is expected in a near future.

The discussion section has been enhanced with a few lines and references:

Similar processes that allow to define IFC-compliant infrastructure entities are a natural future line for this research, once standards such as IFC Road are published and openly available. The feasibility of this line of research is being demonstrated in recent work [Refs].

Reviewer 2 Report

Dear authors and Editor,

I’ve reviewed the paper “3D Point Cloud to BIM: Semi-Automated Framework 3 to Define IFC Alignment Entities from MLS4 Acquired LiDAR Data of Highway Roads”. The paper is useful, suggesting an approach that could be widely used in its field of study, despite some drawbacks that were highlighted by the authors. The introduction provides solid state of the art of road modelling and states clear objectives. The description point by point of the methodology assures the repeatability of the experiment and is very clear and detailed. The subdivision of the Results in sections corresponding to the Methodology ones makes it easier to assess the big number of steps and relative findings. The Discussion section could be improved; the authors stressed some drawbacks of the method; it would be interesting to quantify how much this not complete automation of the process in terms of time. I have only some few specific comments:

Figure 9: adjust the editing (sample right and left edge are written with different styles)

Line 452: does “the values they take for the validation” means the values used for validation?

Lines 462-481: I think these lines, or part of them, should be moved to the Methodology section

Author Response

I’ve reviewed the paper “3D Point Cloud to BIM: Semi-Automated Framework to Define IFC Alignment Entities from MLS Acquired LiDAR Data of Highway Roads”. The paper is useful, suggesting an approach that could be widely used in its field of study, despite some drawbacks that were highlighted by the authors. The introduction provides solid state of the art of road modelling and states clear objectives. The description point by point of the methodology assures the repeatability of the experiment and is very clear and detailed. The subdivision of the Results in sections corresponding to the Methodology ones makes it easier to assess the big number of steps and relative findings.

> Authors want to thank the reviewer for his/her comments, and the effort and time employed reviewing this manuscript.

The Discussion section could be improved; the authors stressed some drawbacks of the method; it would be interesting to quantify how much this not complete automation of the process in terms of time.

> Indeed, an important piece of information that was missing in the original manuscript is a time-based analysis of the manual verification-correction process that is presented in Section 2.2.4. During the validation of the process, the total time of the automatic edge detection process was computed, together with the fraction of the time employed for correcting wrong sections by a manual operator. Authors did not include these data on the original manuscript as the data may not be representative enough to be included in the Results section, but we agree with the reviewer that Discussion section can be enhanced with a few lines that comment these results:

[Section 4: Conversely, it can be argued that the main drawback of this method is the fact that it is not a fully automated process]. This may have an impact on the time consumption of the method. Specifically, for Road Edge Detection (Section 2.2.4), manual corrections were required only for 20% of the point cloud sections but represent more than 65% of the total time of the process. Hence it is clear that full automated and reliable processes would save resources in terms of time and manual interaction.

I have only some few specific comments:

Figure 9: adjust the editing (sample right and left edge are written with different styles)

> Figure 9 has been updated conveniently.

Line 452: does “the values they take for the validation” means the values used for validation?

> Yes. The wording has been modified so it is easier to understand for the reader.

Lines 462-481: I think these lines, or part of them, should be moved to the Methodology section

> Thank you for the comment. Lines 462-481 explain the process and metrics that allow the understanding of results in Tables 3, 4 and Figure 14. Authors believe that moving them to the methodology section would be more inconvenient for the reader, as equivalent content in Lines 498-513 should be moved to the same new section for coherence. This would make an undesired text flow gap between the explanation on how the results are generated and the results themselves.