Authors: Callie Zuck*, Pennsylvania State University
Topics: Spatial Analysis & Modeling, Hazards and Vulnerability, Remote Sensing
Keywords: image analysis, emergency management, OBIA, data gaps, semi-automation, building footprints, data fusion, lidar, orthoimagery
Session Type: Poster
Start / End Time: 1:20 PM / 3:00 PM
Room: Napoleon Foyer/Common St. Corridor, Sheraton, 3rd Floor
Presentation File: No File Uploaded
Emergency management personnel routinely rely on geospatial and remote sensing data when planning for and responding to disasters. One example of this critical geospatial data needed to support emergency management are building footprints. Despite an established need for building footprints many urban areas have large gaps in this essential data layer, hampering emergency planning and response efforts. Existing methods to produce building footprint layers are inefficient (i.e. manual interpretation), and therefore unable to respond to most emergencies and disasters with optimal results. This study presents the creation of a semi-automated building footprint identification method using object-based image analysis to fill in this critical data gap. The method involves using ArcGIS 10.6 and eCognition Developer 9.2 to support object-based imagery analysis. It begins with data fusion of lidar and high-resolution multi-spectral digital orthoimagery as the basis for object-based classification using multiresolution image segmentation and object hierarchy to support ruleset creation in eCognition. The ruleset focuses on spectral and spatial differentiations to classify building footprints. OpenStreetMap contains building footprints for a small portion of the case study area of Bergen County, New Jersey. These building footprints that exist on OSM provide random coverage, are incomplete, and are up to eight years outdated. This study produced a final building footprint layer for the county that was able to update and fill in this critical data gap with increased coverage. The method and workflow presented in this case study can be applied elsewhere to support emergency management.