Skip to Main Content
Visual surveillance systems are often installed in concourses, car park areas and high security sites to monitor the flow of pedestrians and vehicles for security and data analysis. The job of monitoring image sequences is usually assigned to a human operator who waits for important events or specially appointed objects to occur. Operators become bored and lose concentration. It is therefore essential to devise autonomous surveillance systems, which can search and track appointed objects in a wide region and alert a human operator only when appointed objects appear in these watched areas. We report a study on a new region-wide automatic visual search and pursuit surveillance system for appointed objects using networks. By processing image sequences on the basis of object properties, such color and shape (vehicle) or height and gait (pedestrian), the system can search, recognize and track the appointed object. The system uses multiple cameras and every camera carries a set of software. We explain the integrated software system and show an experiment system and experimental results. The system recognizes objects region-wide by using a network. Correct classification yields are 92% for vehicles, 64% for pedestrians. The system is able to process about 2 frames per second.