Detection of Building Equipment from Mobile Laser Scanning Point Clouds Using Reflection Intensity Correction for Detailed BIM Generation

The Building Information Model (BIM) has been increasingly adopted for building maintenance and management. For existing buildings lacking prior digital models, a BIM is often generated from 3D scanned point clouds. In recent years, the automatic construction of simple BIMs comprising major structural elements, such as floors, walls, ceilings, and columns, has become feasible. However, the automated generation of detailed BIMs that incorporate building equipment, such as electrical installations and safety systems, remains a significant challenge, despite their essential role in facility maintenance. This process not only enriches the information content of the BIM but also provides a foundation for evaluating building safety and hazard levels, as well as for supporting evacuation planning and disaster-preparedness simulations. Such equipment is typically attached to ceilings or walls and is difficult to detect due to its small surface area and thin geometric profile. This paper proposes a method for detecting building equipment based on laser reflection intensity, with the objective of facilitating the automatic construction of detailed BIMs from point clouds acquired by mobile laser scanners (MLSs). The proposed approach first corrects the reflection intensity by eliminating the effects of distance and incidence angle using polynomial approximation, thereby normalizing the intensity values for surfaces composed of identical materials. Given that the corrected intensity approximately follows a normal distribution, outliers are extracted as candidate points for building equipment via thresholding. Subsequently, the point cloud is converted into a 2D image representation, and equipment regions are extracted using morphological operations and connected component labeling. Experiments conducted on point clouds of building ceilings and walls demonstrate that the proposed method achieves a high detection accuracy for various types of building equipment.