A novel object tracking algorithm based on discrete wavelet transform and extended kalman filter

A new method for detecting and tracking multiple moving objects based on discrete wavelet transform and Extended Kalman filter is proposed in this paper. Although Kalman filter tracks moving objects accurately, it requires a heavy computational burden. Discrete wavelet transform has a nice property that it can divide a frame into four different frequency bands without loss of the spatial information, we use Kalman filter on low frequency sub-band so it can reduce the computational burden and remove most of the fake motions in the high frequency sub-band. In tracking multiple moving objects, many applications have problems when objects pass across each other. We exploit pattern matching in a simple feature space for solving this problem. The experimental results prove the feasibility and usefulness of the proposed method.