Pixel-Cluster Decomposition Tracking for Multiple IR Sensor Surveillance

Tracking midcourse objects for IR surveillance is a significant problem that must be solved to provide an accurate and consistent set of tracks for intercept and discrimination. Numerica has developed an advanced STSS tracker and tested it extensively on government supplied scenarios. Several key challenges remain when tracking closely spaced objects: (1) some objects remain unresolved through most of midcourse; (2) inconsistent object resolutions may persist over time or between multiple sensor views; and (3) an explosion in the number of observed targets can overwhelm any tracking system with potential targets. To advance the state-of-the-art this program develops signal processing, pixel-cluster tracking, and group-cluster tracking as integrated components that fuse and exploit all available information. The new technology allows us to decompose pixel-clusters into subclusters and resolved objects in support of accurate tracking and early discrimination. Real-time tracking throughout the most challenging scenario scenes is maintained by temporarily resorting to group-cluster tracking. The algorithms are to be fully integrated with Numerica's Multiple Frame Assignment tracking system to make better signal processing and data association decisions by exploiting information from multiple frames of data. This integration provides a rapid transfer of the algorithm into MDA's BMD programs and the commercial market.