A Novel Framework for Tracking Groups of Objects

We describe a novel framework for group tracking and identiication termed ow conservation group tracking. Our framework integrates local kinematic measurements with ow conservation constraints so as to yield a robust estimation of the positions and types of targets across an arena. The framework is based on three components each of which is novel. The measurement-to-track component uses a reduction to minimum cost ow problems. The estimation of the number of objects on each track is based on estimation using ow networks, and the assessment of the types of vehicles is reduced to hierarchy matching algorithms.