You may have heard that PURR may be down temporarily this Thursday (10/17) for maintenance. The maintenance is being rescheduled, and we do not expect to have any downtime this week. We will let you know when the maintenance has been rescheduled. close

Building Information Modelling (BIM) data repository with labels

Listed in Datasets

By Jiansong Zhang1, Jin Wu1

Purdue University

The dataset contains five models with extracted elements for each model. The elements are manually labeled by researchers with construction/civil engineering backgrounds based on discussion and majority vote.

Version 1.0 - published on 09 Oct 2019 doi:10.4231/60V2-PJ72 - cite this Content may change until committed to the archive on 09 Nov 2019

Licensed under CC0 1.0 Universal

Description

The dataset contains five models with extracted elements for each model.

The elements are manually labeled by researchers with construction/civil engineering backgrounds based on discussion and majority vote. Dataset files were created using Autodesk Revit Architecture 2011.

Object signatures have been widely used in object detection and classification. Following a similar idea, the authors developed geometric signatures for architecture, engineering, and construction (AEC) objects such as footings, slabs, walls, beams, and columns. The signatures were developed both scientifically and empirically, by following a data-driven approach based on analysis of collected building information modeling (BIM) data using geometric theories. Rigorous geometric properties and statistical information were included in the developed geometric signatures. To enable an open access to BIM data using these signatures, the authors also initiated a BIM data repository with a preliminary collection of AEC objects and their geometric signatures. The developed geometric signatures were preliminarily tested by a small object classification experiment where 389 object instances from an architectural model were used. A rule-based algorithm developed using all parameter values of 14 features from the geometric signatures of the objects successfully classified 336 object instances into the correct categories of beams, columns, slabs, and walls. This higher than 85% accuracy showed the developed geometric signatures are promising. The collected and processed data were deposited into the Purdue University Research Repository (PURR) for sharing.

Corresponding papers:

Wu, J., and Zhang, J. (2019). “Introducing geometric signatures of architecture, engineering, and construction objects and a new BIM dataset.” Proc., 2019 ASCE International Conference on Computing in Civil Engineering, ASCE, Reston, VA, 264-271.

Wu, J., and Zhang, J. (2019). “New automated BIM object classification method to support BIM interoperability.” J. Comput. Civ. Eng., 33(5), 04019033.

Wu, J., and Zhang, J. (2018). "Automated BIM object classification to support BIM interoperability." Proc., ASCE Construction Research Congress, ASCE, Reston, VA, 706-715.

Cite this work

Researchers should cite this work as follows:

Tags

The Purdue University Research Repository (PURR) is a university core research facility provided by the Purdue University Libraries, the Office of the Executive Vice President for Research and Partnerships, and Information Technology at Purdue (ITaP).