Layered Target Reacquisition and Prosecution from Wide Area Motion Imagery (WAMI) Cues

ABSTRACT: Under this effort, we will develop a system for precise cross-cueing of target locations between different EO sensors flying on different platforms. The primary emphasis is on target handoff between wide area motion imagery (WAMI) sensors and narrow field-of-view full motion video (FMV) sensors. We address the challenges imposed by metadata inaccuracies, different resolutions, different frame rates and different viewpoints by the two sensors as well as dense target environments where the targets of interest may be moving or stationary as well as partially occluded or shadowed. The Phase II project will build upon the Phase I effort which demonstrated that arget handoff is feasible between a wide area EO sensor and a narrow field-of-view EO sensor. In Phase II we will (1) improve the cross sensor registration algorithms to simultaneously take into account multiple frames of data, (2) improve target handoff in dense target environments by incorporating multi-target association using kinematic as well as appearance matching, and (3) expand the feature set used in associating target appearance from one sensor to another to include image correlation, color and feature histograms. In addition, we will integrate the target handoff functions into a reliable and efficient system and comprehensively test system performance and improve robustness and accuracy across a wide range of operating conditions. The system will be tested on measured data from multi-sensor collections as well as simulated data where a broad range of operating conditions may be assessed. The Phase II effort will deliver a target cross-cueing software system and a test report detailing the benefits and limitations of the cross-cueing capabilities relevant to practical applications. BENEFIT: The ability to reliably and accurately cross-cue target information between different sensors operating on different platforms will enable coordination and cooperation between multiple sensor assets to best leverage availability, capability, and coverage of diverse ISR sensors and platforms. Target handoff from one wide area motion imagery (WAMI) sensor to another WAMI sensor or gimbaled full motion video (FMV) sensor enables continuous target tracking beyond the field-of-view of any one single sensor. Future benefits, enabled by this Phase II project, include leveraging of high resolution target features from a narrow field-of-view sensor in the data exploitation processing by a wide area, but lower resolution, sensor to achieve improved situational awareness. Target cross-cueing may also enable the simultaneous exploitation of imagery of the same target from multiple points of view, thus mitigating against illumination artifacts or occlusion effects observed from a particular sensor viewpoint. Furthermore, target cross-cueing may support the integration of information from different sensor modalities in the exploitation process -- e.g., enabling visualization of a target detected by a non-literal sensor.