Recently, an automatic system extensively used for surveillance in wide area environments has been developed to continually track an individual while keep the individual centered in the field of view (FOV). There have long been some knotty problems in the tracking of moving object in computer vision. In a multi-camera surveillance system, both the hand,off and assignment among cameras play an important role in generating an automated and persistent object tracking, a condition typical of most surveillance requirements. There has been a lack of studies in camera hand-off issue in PTZ-camera system. This paper investigates the application of automatic methods for tracking individual across cameras via a surveillance network. We propose a camera hand-off filter which can automatically provide an optimal capacity and a solution to perform the camera assignment efficiently. It comprises three fundamental components. Firstly, each camera determines time to trigger camera handoff process. Then, the execution of camera rotation is applied. Finally, the optimal camera should be selected. Moreover, our approach of individual tracking activates surveillance in cameras and achieves seamless tracking of high quality image. We develop an innovative low computational complexity handoff filter that can automatically carry out the camera assignment and handoff task. In the experiment, with a feasible solution for seamless tracking and real-time surveillance, our algorithm can efficiently and automatically perform the hand-off task in a multiple active camera surveillance systems.
|Number of pages||21|
|Journal||International Journal of Innovative Computing, Information and Control|
|Publication status||Published - 2012 Feb 1|
All Science Journal Classification (ASJC) codes
- Theoretical Computer Science
- Information Systems
- Computational Theory and Mathematics