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

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.本研究提出如下假设：尽管Autodesk Recap Pro与CloudCompare均可为Scan-to-BIM(扫描转建筑信息模型)应用提供满足要求的几何配准精度，但软件处理环境的差异会显著影响作业流程效率、可扩展性以及下游BIM实施性能。 本数据集采集自马来西亚柔佛州马来西亚理工大学土木工程学院结构与材料实验室的钢框架结构，使用Faro Focus Laser Scanner 3D X130(Faro Focus激光扫描仪3D X130)完成采集。 实验数据表明，尽管Autodesk Recap Pro与CloudCompare均实现了适用于Scan-to-BIM应用的可靠几何配准，二者的平均均方根误差(Root Mean Square, RMS)分别为0.85 mm与1.01 mm。但Autodesk Recap Pro在各扫描站位上的均方根误差分布更为稳定；而CloudCompare由于采用点对点ICP(Iterative Closest Point, ICP)配准方法并保留原始点云密度，其均方根误差波动幅度更大。 二者在作业流程效率上也存在显著差异：Autodesk Recap Pro仅需约30分钟即可完成配准流程，而CloudCompare由于需要额外的文件转换步骤与人工干预，耗时约85分钟。此外，Autodesk Recap Pro通过内部过滤与索引操作将最终点云数量减少了26.87%，而CloudCompare则保留了原始采样密度。最终数据体量也存在巨大差异，这体现了两款软件在作业流程优化与最大程度保留数据保真度之间的不同设计理念。 本数据集可作为佐证，证明软件处理架构不仅会显著影响配准精度，还会影响Scan-to-BIM流程中的实际实施效率。Autodesk Recap Pro适用于以施工为导向的作业流程，可缩短处理时长、提升计算可管理性并实现高效的BIM集成；而CloudCompare则在需要详细保留点云数据与开展高级几何分析的研究类应用中仍具有极高价值。 本数据集可供研究人员、工程师与行业从业者用于点云配准软件的对比基准测试、Scan-to-BIM流程验证、数字孪生开发、施工质量控制研究，以及涉及点云数据分割、建模、偏差评估与结构分析的未来研究工作。本数据集还提供了一个可复现的基准框架，用于评估数字化施工环境中配准质量、作业流程效率与下游工程应用之间的关联关系。