Based-on particle filter for vehicle detection and tracking in digital video

In this paper, the visual-tracking problem for the case of maneuvering target is addressed. The environment of maneuvering tracking for the visual information processing is complex often, there is much non-linear motion. The particle filter is applied for this problem. To predict the probably position of the visual-object at the next time by used this method. For initial filter, discrete wavelet transform is used to detect the object, and a suitable model to characterize a state space. By experimental, this method is proved much effectively, and the accuracy of tracking is improved greatly.