Tracking of dynamic objects based on optical flow

In this paper we describe a system, which incorpo-rates the optical flow to navigate in unknown envi-ronments. With the ability to identify and track mov-ing objects it will be able to compute their trajectory and future positions to avoid collisions with them. The system incorporates an algorithm, which elimi-nates the optical flow induced by the motion of the camera (egomotion). Real-world and synthetic data is used in all stages of the development of the system, simulation is used as a primary design aid