Augmented Reality to Enable Users to Identify Deviations for Model Reconciliation

Round 1

Reviewer 1 Report

The research topic chosen by the authors looks interesting and promising, because the issue of integrating various digital tools with BIM of significant practical importance.

The structure of the article is correct, all sections are described logically and correctly. The article is very easy and pleasant to read.

I only have a couple of comments.

1. The article has only a practical orientation without a clear theoretical scientific basis.
2. The authors talk about two advantages of using VR (lines 580 - 591). What exactly is the advantage of using VR in comparison with a real visual inspection performed by a foreman / construction inspector?

Author Response

Please see the attachment

Author Response File: Author Response.docx

Reviewer 2 Report

This is a clear and well written paper describing a study investigating if AR HMDs can support identifying construction deviations from as-designed BIM models. However, there are some weaknesses that need to be addressed before is ready to be published. Please see below detailed comments.

1. The title is a little misleading, the deviation identification and model reconciliation is being done by the user not by AR. AR is only used to overlay a 3D model and the user checks for inconsistencies. Please amend the title to reflect this important difference.
2. Intro, it is not clear how the descriptions about COBIE are relevant to the subject the paper is addressing. Also, strictly speaking COBIE is not a data format. I suggest removing this part or to substantiate it more thoroughly.
3. Please explain briefly the difference between immersive and non-immersive AR. And how this difference might affect the presented use-case.
4. Section 1.1.3 lacks structure and prominence. For example, the BIM section (1.1.1) is as prominent as 1.1.3, even though BIM is only tangentially addressed in the study. Also, there is no discernible structure to the example presented and widely different applications are presented. I would suggest providing more structure to this section based on review works such as: “Research trends and opportunities of augmented reality applications in architecture, engineering, and construction”, “A research agenda for augmented and virtual reality in architecture, engineering and construction”, Identification of application areas for Augmented Reality in industrial construction based on technology suitability”, and “Augmented and Virtual Reality in Construction: drivers and limitations for industry adoption”
5. The big limitation with laser scanning is creating a model out of the point cloud. The most time-consuming task is the modelling not the capturing? Capturing a point cloud enables to automate the process. Can the approach presented here be automated?
6. Line 220-221, Please explain how the model was registered? That is how the model was located in the correct position and orientation?
7. Line 253, Please explain what method was followed and what considerations were taken to incorporate the deviations besides large, mid, and missing. Do these types of deviations usually occur in practice? Are these examples representative of real-life situations?
8. Line 294, please provide more information about the marker, e.g. number of markers used, dimensions, placement on site and orientation. Also, please indicate if a library was used to identify the marker and overlay the model or if it was developed by the authors.
9. Also please include information on how drift problems were addressed and what happens when the camera loses sight of the marker for long periods of time. Some information is provided in line 415, but more detail would be appreciated.
10. How large the test site was? Did the participants needed to walk large distances? I suggest including a diagram of the test site with dimensions and orientations.
11. 502, typo.
12. Major problem with AR is registration and drift, this will impact use in real-conditions dramatically. Also, potential issues with markers were not discussed, e.g. minimum dimensions, etc. It seems that the test site was too small to truly reflect site conditions. For a larger test site many markers might be needed, how will this affect the applicability?
13. This approach requires a rather detailed model, please explain how this requirement will affect applicability in practice.
14. The limited FOV of the HoloLens was not addressed. Please elaborate how this will affect identification. Tablets might work better for this application as it will be easier for the user to see what is there and avoid the encountered problems of identifying objects that are not there. Please explain why the HoloLens was selected rather than a tablet.
15. I suggest including a section that systematically compares this approach with other similar approaches in order to assess the novelty of this paper.
16. The limitations of the presented approach are not addressed thoroughly. For example, this approach will be useful only when a detailed model is available. If not, as the authors described an initial laser scan survey needs to be carried out that defeats the purpose of this approach.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Round 2

Reviewer 2 Report

Most of the comments have been addressed in a sufficient manner to allow this paper for publication.