A Hierarchical Target Tracker Using Image Data

The paper describes a target tracking algorithm for tracking ground based targets using image data. In particular the correlation of ground vehicle orientation (from an imaging sensor) with target velocity (of importance to the tracker) is discussed. It is believed that using this correlation will make the tracker described in this paper superior to trackers that don´t use image data. Furthermore, the tracker proposed in this paper uses terrain map data to help estimate present target position and orientation. This is expected to greatly improve the accuracy of predictions of future target behavior.