“…Given several measures over time, we can use target specific dynamic models to filter the sensor data using a Kalman Filter (KF) [20], an Extended Kalman Filter (EKF) [21], a UKF [22] or a Particle Filter (PF) [23]. PFs in CV were applied to visual tracking tasks such as ball tracking [24], spherical pendulum tracking [25], human face tracking [26], articulated object tracking [27], vehicle tracking [28] or object tracking [29]. The combination of a PF with a UKF, known as a UPF [30], is described in [31] for visual contour tracking, and in [32] for ground maneuvering target tracking.…”