A Particle Filter Based Algorithm for State Estimation of Dim Moving Point Target in IR Image Sequence

摘要

Under the condition of the target's initial information is already estimated successfully, This paper present a realtime target tracking algorithm that based on particle filter (PF) update algorithm. According to the particles' transmission characteristics and the measurements from a single frame detection, the algorithm estimate the target's following moving state, in this way, tracking the dim moving point target successfully under the background of low signal-noise-ratio (SNR) in IR image sequence. The paper introduced the related theory and the concrete steps for accomplishing the algorithm and also simulated the proposed tracking algorithm on the MATLAB platform. Experimental analysis and results showed that the algorithm achieved real-time, dynamic stability and robustness while tracking the random moving target in high clutter environment.