Automatic Deformation Detection and Analysis Visualization of 3D Steel Structures in As-Built Point Clouds

The use of Building Information Modeling (BIM) is a growing reality in the civil industry. Merging 3D geometric information with engineering data, the BIM model combines geometry, spatial relationships, and other properties used at all stages of the building's life cycle. With the need to apply such methodology to existing buildings, researchers have focused on how to automatically generate or update such models from 3D point clouds provided by laser scanning sensors or photogrammetry methods. Most of this research has focused on the recognition of either planar (e.g., floor, walls) or cylindrical (e.g., piping) structures. Few works have dealt with the detection of steel structural elements, due to its particular shape, and only recently focused on detecting their geometric specifications. However, in these approaches, the point cloud of each structural element was manually separated from the point cloud of the entire building. This situation creates a challenge since the manual segmentation of a point cloud is a long and subjective process. In addition to geometric information, recent research has focused on automatically detecting anomalies in structures. Such information, incorporated into the BIM model, allows the structural element to be evaluated using structural analysis applications. The availability of the results of this analysis to BIM professionals is essential for the correct planning of possible interventions. Unfortunately, open-source BIM model visualization tools restrict their functionalities to design review and construction analysis. In this work, we propose an extension of a previous automated method to detect steel elements to identify and measure deformations and a visualization tool that shows the quality of an existing building's structural elements directly in the 3D point cloud. With the results of deformation detection, we compute the quality level of structural elements and present it directly in the 3D view of the building, preserving its spatial context, with the use of colors and annotations.