Peggy Agouris | Department of Earth Systems & Geoinformation Sciences , GMU

Thursday March 6, 2008, 4:30 PM | Research 1 Room 301

Image-Based Automated Extraction and Management of Geospatial Information

This talk will comprise three parts, each one presenting an automated approach towards the solution of a substantial problem in extracting and managing geospatial information from digital imagery.

The first part will address the automated extraction of road networks from high resolution multispectral imagery. The approach that will be discussed presents an innovative methodology for the meaningful exploitation of spectral scene content to enable the full automation of road centerline extraction. Road extraction results with USGS color-infrared images of suburban scene content will be presented to demonstrate the excellent performance of our approach.

The second part will address change detection. An automated approach to integrate object extraction and image-based geospatial change detection will be presented. In this approach the model of deformable contour models (snakes) is extended to function in a differential mode and a new framework is introduced to differentiate change detection from the recording of numerous slightly different versions of objects that may remain unchanged.

The third part will discuss spatiotemporal helixes which are a novel way to model spatiotemporal change. They represent both the movement of an object, as it is expressed by the trajectory of its center, and the changes of its outline. Accordingly, they are highly suitable for communicating the evolution of phenomena as they are captured in sequences of imagery.

More about Peggy Agouris