Multi-camera target relay tracking strategy based on multi-subblock feature matching

In a video surveillance network system constructed by multiple PTZ cameras, when the target access to one camera´s FOV(field of vision) from another´s, it is a key problem in this system to achieve reliable target relay tracking in a large and complex scene. This paper studies a method to determine multi-cameras´ common relay area (RA) and FOV lines, presents a adaptive blocked rule of the tracking target based on target features, and proposes a target gradual handoff and target relay tracking strategy in common RA based on multi-subblock features correlation matching. The emulation experiment shows the efficiency of this algorithm and achieves reliable relay tracking between two cameras.