The project’s main objective consists on developing a cost-effective, intelligent and IoT aware surveillance platform able to detect anomalous incidents in an urban environment; incidents that can be viewed/interpreted as indicators for possible threats to critical facilities, infrastructures and people. In this direction, the technical objectives are to: a) Design and build a novel low-cost embedded framework with local intelligence and real-time alert categorization based on the utilisation of data stemming from various types of IoT sensors (e.g. cameras, CBRN etc.); b) Design and develop a deployment framework that will support the hosting of third-party analytics algorithms (e.g. intrusion detection, bag snatching, etc.); c)Design, implement and deploy a hybrid, Video Surveillance as a Service (VSaaS); d) Develop a visual analytics framework for strengthening decision making mechanisms; and e) Validate and evaluate the proposed platform/solution exploiting both a lab testbed and a pilot deployed inside the campus of NCSR Demokritos.
Currently, a prototype of a low-cost embedded system where USB and IP cameras can forward their data streams for being analysed. Further to that an analytics deployment framework based on the Docker framework has been implemented. As far as it concerns the cloud based VSaaS, a prototype of the back-end system has been implemented. The front-end is still on development. For the visual analytics the skeleton of the system has been designed and a proof of concept prototype has been implemented. Finally, a lab testbed located at FINT’s premises has been implemented for testing and validating the so far implemented prototype systems.