Metric for Automated Detection and Identification of 3D CAD Elements in 3D Scanned Data

Being able to efficiently compare as-built against as-planned 3D states is critical for performing efficient building and infrastructure construction, maintenance, and management. Three-dimensional (3D) laser scanners have the potential to be successfully applied to these tasks. Recent commercial products allow the comparison of 3D scanned and 3D CAD data based on CAD forms. Their current use is however limited due to the large amounts of manual data processing required for extracting useful information. By using 3D Computer Aided Design (CAD) models as representations of 3D specifications and Global Positioning System (GPS) technologies, the authors present an approach for automating the comparison of 3D sensed data and 3D CAD data. This new approach does not perform this data comparison based on CAD forms but on point-clouds.
This paper discusses the fundamental differences between the two approaches, describes the theoretical implementation of the proposed approach, and presents laboratory experimental results confirming the potential impact of the proposed method on industry’s practices.