A PSEUDO 3D motion estimator for moving object estimation in noisy video sequences

This paper presents a three-dimensional motion estimator for use in cases where we have noisy video sequences containing one moving object on a stationary background. The motion estimator is to be used as part of an image enhancing pre-processing step. A Parallel Extended Kalman Filter (PEKF) developed by J. B. Burl is at the heart of this motion estimator, together with additional data mapping techniques for its expansion to what is referred to as a Pseudo 3D motion estimator. This motion estimator is shown with simulations to have much potential for object motion estimation at low image SNR levels, (< 5dB).