Non-Rigidity Objects Tracking in Dynamic Scenes Using Particle Filter

摘要

Particle Filter(PF) method is an efficient tool for non-rigidity objects tracking. The paper presents a Bayesian-based PF method for objects tracking in dynamic scenes. The paper discussed the Bayesian estimation algorithm and the PF process. The color histograms, witch are used as the measurement, corresponding to the particles' rectangles in the sequence images, are compared with the reference histogram to obtain the optimized posteriori probabilities. The robust mean technique is applied to ascertain the objects' positions. Single object and multi objects tracking are performed in the experiments, results from witch are compared with that of. mean-shift algorithm ; thereby the method presented in this paper is proved more efficient.