Linguistic Extraction of Raster Imagery
From MilcordWiki
Overview
Current image extraction processes involve a complex transformation that extracts metadata about the geo-coordinates, georegisters the scene, and extract the texture information and text labels of features and locations. Linguistic Extraction of Raster Imagery (LERIX) represents a new approach for conflation using curvilinear features, road segments and complete road networks, for the registration of maps and imagery.
Need
Imagery analysts cannot keep up with the flood of imagery that must be interpreting and analyzed. Searching imagery collections is difficult as the metadata associated with the files lacks the granularity needed to enable precision searches. Even when the right map is found, geo-registration the image and extracting features and text from the maps is cumbersome.
Approach
Our approach exploits the intelligence implicit in DRG maps as the context for finding and interpreting imagery associated with the area map.
- Extract linear features from images and digital raster graphics (DRG)
- Extract text from raster maps associated with extracted features
- Automatically detect and extract symbol locations
- Conflate vector datasets extracted from satellite images, DRGs, and other geographic information system (GIS) datasets
Benefits
LERIX increases the productivity of image analysts in managing large volumes of imagery.
Applications
The target applications of LERIX include GEOINT (Geo Intelligence) programs at NGA, Air Force, Army, the Marines, and DHS.
References
- Doucette, P., Kovalerchuk, B., Brigantic, R., Seedahmed, G., Graff, B. "A Methodology for Automated Vector-to-Image Registration" Applied Imagery Pattern Recognition Workshop, 2007. AIPR 2007. 36th IEEE Volume , Issue , 10-12 Oct. 2007.
- Windholz, T., Caglayan, A. and Brigantic, R. T. (2007) “LERIX: Language Enabled Raster Imagery Exploitation”, Technical Report, DTIC AD No. ADB331211, Milcord LLC, Waltham, MA July 2007.
