Beyond Crude 3D Models: From Point Clouds to Historical Building Information Modeling via NURBS

Abstract

This paper presents a complete methodology for HBIM generation based on NURBS derived from a set of point clouds. NURBS combined with profiles and anchor points allow one to model the shape of complex historical objects surveyed with laser scanning techniques. As NURBS are mathematical functions defined through mathematical vectors, they can be converted into BIM objects in order to add semantic information. 1 Introduction Nowadays there is an incredible interest in Building Information Modeling (BIM). BIM technology moves building design from standard 2D (plans, sections, etc.) or 3D representations (X,Y,Z information) towards a new n-dimensional concept based on additional parameters (e.g. time, costs, etc.) that become an effective part of the design phase [1-3]. The generation of an accurate BIM has a direct impact on the methodology for 3D modeling that cannot be based on the creation of predefined static shapes, solids (planes, spheres, cylinders, and so on) or mesh models. Parametric building modeling gives a new dynamic design concept where objects can be edited and modified without redrawing: the shape is connected to a (finite) number of geometric quantities that can be numerically modified. However, BIM is much more than geometry. It is a dynamic database of the building with an improved coordination of construction documents where geometry, spatial relationships, geographic information, and other quantities or properties of building components are structured. BIM technology is one of the most promising recent developments in the architecture, engineering, and construction industry. BIM models can be used for several purposes,