A Particle Filter Algorithm for Target Tracking in Images

We present in this paper a new algorithm for target tracking in cluttered image sequences using the bootstrap particle fil- ter. The proposed algorithm incorporates the models for target sig- nature, target motion and clutter correlation and allows for direct tracking from the image sequence. Monte Carlo simulation results show that the bootstrap tracker outperforms the association of a sin- gle frame maximum likelihood position estimator and a Kalman-Bucy filter (KBf) in a scenario with a heavily cluttered, dim target.