LivingBIM: As-designed and As-built 3D Models and Point Clouds for an Academic Building

LivingBIM is a dataset of a single, recently constructed academic building. It includes as-designed BIM models, as-built 3D point clouds, semantically segmented point clouds, and panoramic images. The dataset was created by Abdullah Alsuhaibani at The University of Texas at Austin and specifically includes mechanical and electrical rooms to represent geometric complexities.

Use Cases

• Train models for semantic segmentation of 3D point clouds based on the provided segmented data.
• Develop algorithms for aligning as-designed BIM models with as-built point cloud data.
• Create virtual tours or digital twins of facilities using the provided panoramic images and 3D models.
• Study geometric complexities in mechanical and electrical systems within AEC projects.

Strengths

• Includes both as-designed (BIM) and as-built (point cloud) representations of the same facility.
• Contains semantically segmented 3D point clouds, which provide ground truth for machine learning tasks.
• Covers geometrically complex areas like mechanical and electrical rooms, which are often challenging for modeling.

Limitations

• Row count and dataset size are unknown, which may limit suitability assessment for large-scale training.
• Column-level documentation is absent; field semantics must be inferred after download.
• Last update date is unknown; freshness unverified.

Provenance

Source

The University of Texas at Austin

Collection Method

Likely collected via 3D laser scanning and BIM modeling of a recently constructed academic building.