Distributed Implementation of Track-Before-Detect

Data Fusion Corporation and Space Computer Corporation proposes a distributed-sensor version of the track-before-detect method for detecting faint moving objects in noisy, cluttered images. Track-before-detect combines multiple images from a single sensor to detect a moving object's trajectory as a whole, versus detecting single snapshots. Fusion of distributed sensor observations improves detectability and reduces false alarms using consistency checks. Communications bandwidth in the distributed system is kept low by transmitting only detection information, not raw data. The result is higher detection and lower false alarm rates via a feasible system architecture with lower sensor and feasible communication requirements. While intended for SDIO's Brillant Pebbles and Brillant Eyes sensors and scenarios, other remote activities will benefit: target recognition, environmental monitoring, and industrial, geophysical, and medical measurements.