Algorithm for beam deformation modeling from LiDAR data

• A methodology for measuring and modeling deformed beams from TLS data was developed.
• Deformations in real beams may be critical in as-built modeling in terms of safety.
• The values of beam deformation are compared with the values of the building codes.
• The end product is a 3D parametric model of the deformed beam defined by two curves.

This paper presents an approach for 3D modeling of beams in structures using laser scanning after identifying several deficiencies in current methodologies of model extrusion. It highlights the errors due to the actual deformation of beams, which does not correspond with the idealized straight line. To avoid these errors, a novel algorithm has been developed. This algorithm performed the segmentation of a beam point cloud by means of a 2.5D approach based on digital image processing. Thereafter, a polynomial surface fitting of flange points is performed in order to model the deformations of beams based on the combined effects of bending moment and torsions deflections. The tool automatically compares the idealized straight plane of flanges with a 3D curved surface reconstructed from the scanned data and analyzes whether it is above the maximum limits established by structural codes. Finally, the end product is a 3D model of the deformed beam.