Enhanced strong Kalman filter applied in precise video tracking for fast mobile target

To tackle the Kalman filter's performance degradation caused by unknown process uncertainties, the author proposes an approach of using strong Kalman filter to effectively implement the precise video tracking for the target with high mobility. An on-line calculated time-variant recession matrix is induced into the predicted estimate error covariance matrix to adjust the Kalman gain, making the gain self-adaptively suit to the innovation sequence, thus resulting in better filtering performance. The strong Kalman filter is enhanced with time-variant recession matrix improved in this paper. The proposed scheme is put to test suggesting that the enhanced strong Kalman filter is capable of tracking the fast mobile target with higher precision than generic Kalman filter.