Cyber System:Collaborative Research: An Intelligent Network of Wireless Videos for Dynamic Scene Analysis

点击翻译按钮获取中文摘要

This proposal focuses on developing an intelligent network of video cameras connected over a wireless communication channel for applications in dynamic scene understanding. Existing video sensor network systems employed in various applications usually comprise of a set of video cameras that transmit data over a wired communication channel to a central processing unit with almost no intelligent processing of the video locally at the senor. The lack of mobility and difficulty of installation of wired communication is a serious impediment to deriving the maximum possible benefit from large networks of video data.Intellectual Merit: Realization of this goal depends on successfully addressing a number of theoretical and practical challenges, including self-organization of camera networks and their operation, learning behavior patterns for scene analysis, optimal utilization of available resources, handling dynamic topology and varying degrees of mobility, distributed control and decision making, and developing computationally efficient algorithms, among others. The proposed research addresses these challenges by developing a hierarchical architecture of a wireless network of video cameras with local intelligent processing at each camera node.Broader Impact: Wireless video sensor networks which integrate image processing, computer vision and networking tasks will help develop a new, exciting and challenging scientific research area at the intersection of video and network technologies. It will immensely benefit a large number of existing applications and create new applications ranging from national priority areas like homeland security, monitoring nuclear installations and border security, to commercial interests like video communications and multimedia entertainment. These PIs will use this opportunity to develop inter-disciplinary courses on wireless video networks that can be taught at both institutions through video-based distance learning technologies.