Due to the increasing threat posed by crime, industrial espionage and even terrorism, video surveillance systems have become more important and powerful during the last years. While most commercially available surveillance systems have to be managed by human operators, who constantly monitor all video streams, several experimental systems from different research groups already include robust video processing approaches for (semi-) automated surveillance. Still, most research activities focus on a sensororiented approach to video analytics of large and distributed camera networks, aiming to extract, analyze and store all extractable information from the video streams. In real-life applications, however, only a limited set of specific threats needs to be covered. Accordingly, only a small subset of potentially extractable information has to be monitored. Besides the huge amount of raw video data, modern surveillance systems are also extended with other sensors that deliver even more data. As a consequence, a new paradigm is introduced, called task-oriented information and data processing for surveillance systems. In the proposed system NEST (Network Enabled Surveillance and Tracking) following the task-oriented approach, every resource allocation, data acquisition, and analysis process is assigned to a specific surveillance task. In order to meet the requirements of taskoriented surveillance, the proposed architecture combines a Service-Oriented Architecture with an Event-Driven Architecture (Event-driven SOA).

No abstract

Real-time object tracking, feature assessment and classification based on video are an enabling technology for improving situation awareness of human operators as well as for automated recognition of critical situations. To bridge the gap between video signal-processing output and spatio-temporal analysis of object behavior at the semantic level, a generic and sensor-independent object representation is necessary. However, in the case of public and corporate video surveillance, centralized storage of aggregated data leads to privacy violations. This article explains how a centralized object representation, complying with the Fair Information Practice Principles (FIP) privacy constraints, can be implemented for a video surveillance system.

Modern surveillance systems collect a massive amount of data. In contrast to conventional systems that store raw sensor material, modern systems take advantage of smart sensors and improvements in image processing. They extract relevant information about the observed objects of interest, which is then stored and processed during the surveillance process. Such high-level information is, e.g., used for situation analysis and can be processed in different surveillance tasks. Modern systems have become powerful, can potentially collect all kind of user information and make it available to any surveillance task. Hence, direct access to the collected high-level data must be prevented. Multiple approaches for anonymization exist, but they do not consider the special requirements of surveillance tasks. This work examines and evaluates existing metrics for anonymization and approaches for anonymization. Even though all kinds of data can be collected, position data is still the one with the highest demand. Hence, this work focuses on the anonymization of position data and proposes an algorithm that fulfills the requirements for anonymization in surveillance.

During the last decades surveillance systems developed from analog one camera one monitor systems to highly complex distributed systems with heterogeneous sensors that can handle surveillance tasks autonomously. With raising power and complexity, ensuring privacy became a key challenge. An event-driven SOA architecture that follows the privacy by design principle is a promising approach to realize a smart surveillance system and is described in this work. Enforcement of privacy is not only complex for engineers and system designers; rather it is not understandable for the average user, who cannot even assess potentials and limitations of smart surveillance systems. This work presents an approach for privacy that is focused on the user, i.e., the observed subject. By using a mobile device the user can interact with the surveillance system and is not passive anymore, as in conventional surveillance deployments. This restores the balance between the observed and the observers, enhances transparency and will raise the acceptance of surveillance technology. In the highlighted approach the user can control his individual-related data and privacy preferences and can use services that are benecial for him.