Point Cloud Registration Dataset for Comparative Scan-to-BIM Assessment Using Autodesk ReCap Pro and CloudCompare

Description

This research hypothesis that although Autodesk Recap Pro and CloudCompare are both capable of producing acceptable geometric registration accuracy for Scan-to-BIM applications, differences in software processing environments significantly influence operational workflow efficiency, scalability, and downstream BIM implementation performance. The dataset was generated from Faro Focus Laser Scanner 3D X130 conducted on steel structural frame located at the Structure and Materials Laboratory, Faculty of Civil Engineering, Universiti Teknologi Malaysia, Johor, Malaysia. The data demonstrated that although both Autodesk Recap Pro and CloudCompare achieved reliable geometric alignment suitable for Scan-to-BIM applications, with mean RMS of 0.85 mm and 1.01 mm, respectively. However, Autodesk Recap Pro produced more stable RMS distributions across scan positions, while CloudCompare exhibit greater RMS fluctuations due to its point-to-point ICP registration approach and preservation of original point cloud density. Significant differences were also identified in workflow efficiency. Autodesk Recap Pro completed the registration workflow in approximately 30 minutes, whereas CloudCompare required approximately 85 minuted due to additional file conversion and operator intervention. Furthermore, Autodesk Recap Pro reduced the final point cloud count by 26.87% through internal filtering and indexing, while CloudCompare preserved the original sampling density. Large differences in final data volume were also observed, demonstrating contrasting software philosophies between workflow optimization and maximum data fidelity preservation. The dataset can be interpreted as evidence that software processing architecture significantly affects not only registration accuracy but also practical implementation efficiency in Scan-to-BIM workflows. Autodesk Recap Pro demonstrated advantages for construction-oriented workflow requiring reduced processing time, improved computational manageability, and efficient BIM integration, whereas CloudCompare remains highly valuable for research-oriented applications requiring detailed point cloud preservation and advanced geometric analysis. The dataset may be used by researchers, engineers, and practitioners for comparative benchmarking of point cloud registration software, validation of Scan-to-BIM workflows, digital twin development, construction quality control studies, and future research involving segmentation, modeling, deviation assessment, and structural analysis from point cloud data. The dataset also provides a reproducible benchmark framework for evaluating the relationship between registration quality, workflow efficiency, and downstream engineering applications in digital construction environments.

Categories

Funders

• University of Technology Malaysia

  Johor Bahru

  Grant ID: Research Grant Hi-Tech (F4+), Q.J130000.4622.00Q65.

Licence