SV3D · Surveillance platform based on multi-source video analytics, localized data and cognitive interfaces""

The current surveillance systems used in big buildings or infrastructures are based on the juxtaposed display of multiple alarms, sensors, and independently captured and processed video feeds. That array of inputs generates such a volume of information that is near to impossible to manage by the system users. The SV3D project aims at designing and implementing a security software platform to face these issues.To overcome this usability problems, cognitive and perceptual factors will be key factors in the design and evaluation phases of the project. They will guaranty improvements in the quality of work of operators in a control room, and how it will help them to avoid usual issues like tiredness, relaxation or dispersion in front of multi-channel stimulus overload. One of the solutions envisioned is the use of tridimensional computer generated environments to navigate the surveilled scenarios in a more comprehensive way integrating all aspects of video surveillance, such as sensing, alarm detection, etc.Another main novelty lies in the exploitation of the robust and accurate tracking capabilities offered by multi-cameraenvironments, where several cameras cover the same scene from different viewpoints. Considering the surveillance problem in a multi-camera environment not only mitigates the difficulty of scene understanding (detection and tracking) caused by reflection, occlusion and shadow in the single view case, but also offers higher flexibility in producing relevant video reports.The system architecture will be open and flexible, scalable and distributed, and all communication protocols used in theproject, will be based on standards (e.g. ONVIF), making it compatible with the widely installed legacy systems and leaving the whole data ready to be exploited by intelligent analysis systems to improve the precision of the real time alarm triggering or the after-incident forensic procedures.""