Intelligent video surveillance
The NEST (Network Enabled Surveillance and Tracking) project specialises in finding a solution to the following problem: Today's situation requires a modern surveillance system due to increasing crime, terrorism and industrial espionage. Most video surveillance systems are still manually controlled by human "operators". However, they can only manage a limited number of monitors or video sensors and their attention usually diminishes after a short time.
The Fraunhofer IOSB is developing a new generation automatic surveillance system - the IOSB NEST System - as part of its own research project NEST. Defined surveillance tasks can be carried out autonomously by this system and the security personnel receive automated support in assessing the situation. The system takes over central tasks, such as tracking suspicious persons and visualises the data in a situation picture. This allows the security officer to concentrate more on higher-value tasks, e.g. situation assessment.
Scenario
Karlsruhe, Monday morning. A guest enters the foyer of the Fraunhofer IOSB and announces his visit to Mr. Mustermann at the reception. The reception staff enters the name and destination of the visit in the NEST user software. Then the receptionist asks the guest to look into the camera above the reception counter and seconds later a visitor's badge with a photo is printed out.
The visitor makes his way to Mr. Mustermann and the NEST system takes up the assistance activity in the background. A route planner automatically calculates the possible routes from the foyer to Mr. Mustermann's office. Cameras keep the guest in view. A NEST service checks cyclically whether the guest is on the right path and whether he is in permitted areas of the building. If not - and only in this case - the staff at reception is notified. If the guest has arrived at Mr. Mustermann's, the monitoring task is acknowledged with a message to the reception staff.
Solution approaches
With a dynamic, service-oriented architecture, processes are directly linked to objects or monitoring tasks and not to sensors. This ensures that the tasks can also be performed in very large sensor networks with minimal computing load. The computing capacity is not limited by the number of sensor nodes, but by the number of tasks to be processed simultaneously. Within the framework of the NEST research project, this concept will then prove its applicability.