Classifying and Georeferencing Indoor Point Clouds With ArcGIS

Abstract

This study aimed to develop and apply a manual procedure for clas-sifying and georeferencing indoor point clouds that we created using Paracosm's PX-80 handheld three-dimensional laser scanner. We collected data for 11 buildings in Connecticut, USA and focused on classifying features-of-interest to public safety personnel (i.e., doors, windows, fire alarms, etc.). ArcGIS Desktop was used to manually digitize features that were easily identified in the point cloud and Paracosm's Retrace was used to digitize small features for which imag-ery was needed for identification. We developed several tools in Python to facilitate point cloud classification and georeferencing. The procedure allowed accurate mapping of features as small as a sprinkler head. Point cloud classification and georeferencing for a 14 000 m2 building took 20–40 hours, depending on building characteristics and the types of features mapped. The methods can be applied in map-ping a wide variety of features in indoor or outdoor environments.