Moving target recognition and tracking algorithm based on multi-source information perception

This paper proposed a non- monitoring video multi- object tracking algorithm based on fast resampling particle wave filtering in order to improve the unsupervised monitoring effect of video moving objects. Each particle (sample) can represent the real state assumption. In each time step, the likelihood function is used to evaluate and quantify each particle. The estimated state is approximated by the average value of all the particles after evaluating each particle. To avoid degradation, we use particle resampling to create a new weighing set. Finally, by the simulation test of a real monitoring image show that the proposed algorithm improves the tracking accuracy by more than 20% and the computational efficiency by more than 30% compared with the contrast algorithm, which verifies the effectiveness of the proposed method.