This paper proposes an algorithm for managing redundant measurements provided by a stereo multi-camera system to achieve accurate visual tracking of a moving object. Selfocclusion, different visual resolution zones and optimal selection of redundant measurements are some of the problems addressed. The algorithm uses the extended Kalman filter embedded in a computational efficient pose estimation procedure based on Binary Space Partitioning Tree geometric modelling of 3D objects. Experimental results are presented for the case of an object moving in the visual space of two fixed cameras.

A method of data segmentation, based upon robust least K-th order statistical model fitting (LKS), is proposed and applied to image motion and range data segmentation. The estimation method differs from other approaches using versions of LKS in a number of important ways. Firstly, the value of K is not determined by a complex optimization routine. Secondly, having chosen a fit, the estimation of scale of the noise is not based upon the K-th order statistic of the residuals. Other aspects of the full segmentation scheme include the use of segment contiguity to: (a) reduce the number of random sample fits used in the LKS stage, and (b) to “fill-in” holes caused by isolated miss-classified data.