Motion estimation in flotation froth using the Kalman filter

Machine vision systems have been used to monitor mineral froth flotation systems since the 1990s and their ability to track key performance indicators of the systems online is critical to improved plant operation. One of the challenges faces by these computer vision systems, is estimation of the motion of the froth, which is hindered by the simultaneous deformation, bursting and merging of bubbles. In this paper, we propose a block based motion estimation method using Kalman filtering to improve the motion vector estimates resulting from the new-three-step-search technique. Experimental results derived from flotation froth video sequences are presented.