Motion Estimation for Improved Target Tracking with a Network of Cameras

摘要

Camera systems are an essential tool in security monitoring systems. Typically, the video is either recorded and archived for review after a security breach, or is actively monitored by security personnel for a real-time or preemptive response. Prompted by human factors and manpower issues related to real-time security monitoring for wide area survellience, some automation exists that can monitor the video and alert personnel of a change in the video. Yet, a need persists to automatically monitor targets of interest, in the presence of constantly changing video with similar targets in the scene. Preliminary efforts in this work provide a method to automate video monitoring for security applications through a novel motion estimation method.