Semi-Automated Processing and Incorporation of 3D Geo-Coding Extensions to Commercial Buildings for Emergency Response Applications

• NCJ Number: 237987
• Author(s): Bill Ribarsky; Kalpathi Subramanian; Jianfei Liu; Kyle Lyons; Onyewichi Obirieze
• Date Published: March 2009
• Length: 10 pages
• Grant Number(s): 2007-DE-BX-K010
• Agencies: NIJ-Sponsored
• Publication Type: Program/Project Description
• Award: Three Dimensional Extensions to Geo-Coding Technology for Emergency Response Applications

Annotation

This report describes new automation tools for processing two-dimensional data on a commercial building's geometry, so as to facilitate the visualizing of the building's layout, which is important in planning a law enforcement response to a critical incident in the building.

Abstract

The report presents new tools for automatically processing building data from their original two-dimensional (2D) CAD (computer aided design) files into a three-dimensional (3D) graph representation. This involves four steps. The first step involves data acquisition/preprocessing. This step starts with the raw CAD files, georeferences the data to a base map, and does ensures that data are accurate and consistent. The second step involves the creation of an "adjacency graph," which consists of deriving spatial adjacency relationships from the raw polygon data in the CAD files, resulting in a 2D graph for each floor of the building. The third step pertains to "centerline extraction." In this step, a distance transform-based algorithm is used to extract automatically the centerline (medial axis) of the hallway polygons that typically compose the evacuation routes for the building. The final step involves the "building graph construction." In this step, the adjacency graph and extracted centerline are used to determine the rooms adjacent to the centerline and build a 2D network for each floor. These 2D networks are then combined with the stairways, elevators, and building entrances, resulting in the 3D building network. This report also describes initial work on extending the system to a mobile platform. In addition, work is reported on integrating an evacuation model into the framework. The proposed methodology has been used to process 22 academic buildings at the University of North Carolina at Charlotte. 9 figures, 2 tables, and 24 references