Representation Requirements of As-Is Building Information Models Generated from Laser Scanned Point Cloud Data

Obtaining and utilizing as-is information with 3D imaging technologies, such as laser scanners, during various phases of facility life-cycle is becoming common practice. Both during construction and facility operations information derived from laser scanner data can serve many purposes by providing accurate information about the conditions of the facilities at the time of the scanning. Currently, the information derived from point clouds is typically represented as building information models (BIMs). Despite the benefits of having accurate as-is BIMs, current BIM approaches and tools have limitations in representing as-is information. The reason is partly that current methods of as-is BIM generation are based on existing as-designed BIM generation and representation processes. We identified five main concepts that are unique to as-is BIMs and are not represented with existing BIMs. These characteristics are point density, noise, occlusions, model deviations, and the links between the points and building components.This paper investigates these unique characteristics of as-is conditions and discusses how representing them within a BIM can provide advantages to downstream users, such as enabling decisions based on more complete data than would otherwise be possible.