AFOS: Automated Geopositional Feature-on-Video Overlay Service Navy SBIR FY2010.1 Sol No.: Navy SBIR FY2010.1 Topic No.: N101-100 Topic Title: AFOS: Automated Geopositional Feature-on-Video Overlay Service Proposal No.: N101-100-1443 Firm: Mosaic ATM, Inc. 801 Sycolin Road Suite 306 Leesburg, Virginia 20175 Contact: Stephen Pledgie Phone: (540) 589-2863 Web Site: http://www.mosaicatm.com Abstract: Mosaic ATM proposes to develop an Automated Geopositional Feature-on-Video Overlay Service (AFOS) that provides the capability to accurately superimpose (layer) geospatial information on top of live aerial video feeds. System execution will employ Detail and Resource Throttling (DaRT) to achieve high overlay configurability and adaptation of processing to accommodate changing optical / viewing conditions and resource availability within a Service Oriented Architecture (SOA) environment. We compute and refine the pose of an optical sensor using a multi-rate iterated filter whose measurements are dynamically configured online as a function of scene geometry, video quality, and processing residuals. These capabilities enable the AFOS system to achieve a high-degree of fully-automated performance, delivering persistent quality of service to the human operator across a spectrum of environmental conditions. Benefits: Mosaic ATM's AFOS system for geopositional features and vector data will open a new era for deriving situational awareness from video feeds, allowing users to more effectively leverage historical data stores when reviewing live video streams for objects of interest. Human operators will be able to rapidly orient themselves by viewing a single image stream as opposed to requiring multiple video feeds and handheld maps. AFOS will support multi-person collaborative planning and execution of military operations by enabling multiple operators to view a common information-enhanced video feed. Beyond DoD applications, AFOS will enable first-responders to relay feature-rich video to follow-on personnel, preparing them for the mission ahead and generating a powerful forensic video tool that fuses visual information with geospatially-defined features and dynamically-updated interests points. Return