Tracking Using Peer-to-Peer Smart Infrared Cameras
Final rept. 15 Aug 2007-14 Sep 2008
GEORGIA TECH APPLIED RESEARCH CORP ATLANTA GA
Pagination or Media Count:
This project has two complementary goals to develop new methods for video tracking using multiple, physically distributed cameras without relying on centralized servers and to develop new algorithms for tracking using infrared images. We must develop target models for infrared that are robust to changes in target orientation, illumination, and environmental conditions. In general, the best results are obtained by fusing data fusion of multiple channels, such as visible and infrared and fusion of estimates from the physically distributed nodes in the system. We are particularly interested in physically separated cameras- -practical systems will not always have a visible camera and an infrared camera paired at every node.