A new surveillance system developed by a University of Rhode Island researcher automatically tracks moving objects in real time.

"This camera has broad impact for security surveillance because it eliminates having a full-time guard watching a video screen," says Ying Sun, URI professor of electrical engineering. "It's one intelligence level above any other existing system, and we've found the right compromise between speed and accuracy," he adds.

The camera analyzes images for any motion, at a rate of 15 frames/sec. The technology is based on an image-processing algorithm for real-time tracking. The effectiveness and computational efficiency of the algorithm lets the feedback control loop quickly track motion. The algorithm is implemented in Visual C++ on a PC. However, it can also be configured to operate on an embedded PC, handheld computer, or DSP. Motion can trigger video recording with the resulting images stored on a hard disk as AVI files.

The system is also inexpensive. It can operate on a $30 Webcam and requires a motor-driven, pan-tilt camera mount and a processor. With inexpensive equipment, the system could cost under $300."We're working on adding 'behavior modifiers' as well, so that once the camera identifies motion it can be programmed to continue tracking a given size, shape, or color, regardless of any other motion," says Sun.

The system can provide property surveillance at ATM machines, businesses, warehouses, factories, and homes, as well as handle homeland defense and military use.