A two‐stage recursive ray tracing algorithm to automatically identify external building objects in building information models

Abstract

The internal or external attribute of building information modeling (BIM) objects is vital information for many BIM‐based engineering analyses such as building energy analysis and cost estimation. Unfortunately, such information is often inaccurate, incomplete, or missing entirely in most BIM models. Manually checking, correcting, or inputting this data for large‐scale BIM models can be time‐consuming, laborious, and error‐prone. This study proposes a two‐stage ray tracing algorithm to automatically identify external BIM objects, based on the idea that external objects of a building can be viewed from somewhere outside the building. The first‐stage ray tracing samples viewpoints from the six faces of an offset axis‐aligned bounding box (AABB) of the building and emits a ray for each viewpoint to detect relevant external objects. The second‐stage ray tracing recursively searches for any remaining external objects from the view of the external objects that have been detected in the previous round of ray tracing. Both stages are carefully designed for efficiency. Furthermore, a two‐tier AABB tree is introduced to spatially index building objects on both model and object levels to accelerate relevant geometry operations. The proposed algorithm is validated with one synthetic and two large‐scale real‐world building models. The results show that all the external objects in the three models are accurately and efficiently identified, and the two‐tier spatial indexing and other acceleration techniques improve the algorithm's efficiency significantly.

Iterative application of generative adversarial networks for improved buried pipe detection from images obtained by ground‐penetrating radar

“…Object detection is also used for various purposes in civil engineering. For example, there are studies on detecting vehicles, pedestrians, and cyclists for autonomous driving (Carranza-García et al, 2022); detecting workers and heavy equipment for work zone safety (Shen et al, 2021); detecting external building objects to improve the accuracy and completeness of bridge information modeling (BIM) data (Ying et al, 2022); detecting the floating part from infrared images (Chun & Hayashi, 2021); detecting damaged parts from a 3D bridge model (Yamane et al, 2022(Yamane et al, , 2023; detecting material types of object surfaces for effective disinfection by robots (Hu & Li, 2022); and detecting unexploded ordnance materials in shallow waters to remove potentially dangerous objects (Foresti & Scagnetto, 2022).…”

A deep natural language processing‐based method for ontology learning of project‐specific properties from building information models

“…The conversion of internal data models of various software to an IFC data file needs to map different levels of information related to element geometry and semantics. Even though object types and geometries can be accurately converted into IFC‐compliant models in most systems, mapping built‐in object properties with IFC‐predefined properties and property sets is often inadequate (Ying et al., 2022). For example, Autodesk Revit allows some basic parameters, such as “level” or “height,” automatically mapped to corresponding IFC properties.…”