Geospatial Exploitation (GEO)
Abstract
(U) The Geospatial Exploitation (GEO) thrust will provide a new set of geospatial intelligence (GEOINT) products, continuously updated and maintained in a form that ensures their consistency across both product elements (digital elevation models, traditional maps, 3-D structure models, census summaries, and directories) and spatial nodes (coarse resolution country data for economic analysis to fine resolution building data for platoon-level combat operations). Techniques of interest include model-based image analysis (both object recognizers and change detectors), symbolic correlators (both temporal and spatial), and emerging cognitive methods to identify changes to objects, addresses, names, and functions of natural and human-made structures. These algorithms will be scaled to operate on data streams including full-motion video, Laser Identification Detection and Ranging (LIDAR), multi- and hyper-spectral, synthetic aperture radar (SAR), and Geographic Information Systems (GIS) in addition to conventional electro-optical (EO) geospatial imagery. GEO algorithm architectures will be explored to achieve scalability through spatial, temporal and ontological partitioning. GEO technologies are planned for transition to the National Geospatial-Intelligence Agency (NGA). Activities funded within the GEO research space include: • The Urban Reasoning and Geospatial Exploitation Technology (URGENT) program is developing a 3-D urban object recognition and exploitation system that enables advanced mission planning and situation analysis capabilities for the warfighter operating in urban environments. URGENT will create techniques for the rapid exploitation of EO and LIDAR sensor data at the city scale to recognize urban objects down to the soldier scale. URGENT will apply image processing technology to geospatially registered 2-D/3-D data collected from airborne and terrestrial sources, yielding precise annotations for the objects in an urban area. URGENT will also develop a 3-D reasoning engine to query object shapes, locations, and classifications for advanced geospatial exploitation capabilities. • The Geospatial Representation Integrated Dataspace (GRID) program is developing an automated geospatial data fusion, modeling, and dissemination system from national assets for the tactical warfighter. Geospatial registration algorithms will automatically fuse geospatial data from multiple sources including EO, LIDAR, SAR, and hyperspectral - and encode the fused data as a temporally indexed volumetric model that drastically reduces geospatial data storage requirements while enhancing image quality. Updates will propagate to the model using a compressed geospatial data format capable of reaching the warfighter even with the bandwidth constraints of tactical networks.
Document Details
- Document Type
- Accomplishment
- Publication Date
- Oct 01, 2011
- Source ID
- 353f6f9a444f1af869e21853709c3985