Algorithms for IR Data

ABSTRACT:Toyon Research Corporation proposes the research and development of advanced algorithms and efficient software for the detection, tracking, and geo-location of dim targets in infrared sensor data. The algorithms will provide detection of both moving and stationary targets with novel, temporal signatures based on near-optimal exploitation of high-frame-rate satellite-based IR imagery. The proposed detection and tracking algorithms are based on innovative development in a Bayesian track-before-detect framework. The geo-location algorithms are based on the coupling of mission image-to-reference frame alignment using a robust statistical algorithm, frame-to-frame feature tracking and registration, and estimation of physical sensor parameters to remove geo-registration biases. In addition, the utilization of advanced clutter modeling and rejection processes allows for the capability of detecting targets with intensities several orders of magnitude below image clutter. The developed algorithms will be implemented in efficient software libraries and delivered to the government for use in IR data exploitation processes. A prototype software application will also be developed and delivered for end-to-end demonstration of the IR sensor processing system. Toyon will provide software integration support and will perform extensive testing of all developed algorithms and software using real-world high-frame-rate IR sensor data, as well as simulated data to address algorithm limits and government interests.BENEFIT:The successful completion of this R&D will result in advanced algorithms and a real-time software implementation for detection, tracking, and geo-location of dim targets (moving and stationary) from space-based IR sensor platforms. A direct application will be insertion of the algorithm suite into existing baseline OPIR software at Air Force ground stations and/or technical intelligence labs, such as the Aerospace Fusion Center, to be used to process image data from existing sensors. The benefits of this successful R&D would be a comprehensive dim target detector to aid the Air Force in finding and locating adversarial threats, and locating threat development and test sites. This software could also be optimized for programs that use satellite sensors for surveillance such as NASIC ATEP II or the AFRL EAGLE HTI Space Experiment.