A track-before-detect algorithm based on particle filter with model estimation

An optimal nonlinear Bayesian TBD (track-before-detect) algorithm based on particle filter for dim targets detection and tracking in cluttered background is proposed. To make full use of the targets a priori information, a table of possible target movements and their transition probabilities among these base states are introduced. Furthermore, the TBD technology is exploited for the target detection in cluttered image sequences with low SNR. Here particle filter is provided to implement the Bayesian regression and estimate the target´s state model at each step.