MEC PoC#8 Webinar – Video Analytics

As urbanization continues, cities need to become smarter, which includes leveraging information systems for improving the safety and wellbeing of their citizens. Many of these systems rely on extensive deployment of distributed sensors. These sensors generate large amounts data which calls - in particular in the context of public safety - for efficient real-time processing strategies in order to detect incidents as quickly as possible for initiating appropriate counter actions. Also, sensor data handling is typically governed by tight requirements for data privacy, meaning that data needs to be appropriately protected and cannot be carried outside of the city perimeter. Video Analytics is an end to end use case for providing video surveillance to cities, municipalities, and enterprises over an LTE network. Mobile Edge Computing (MEC) is used for analysing raw video streams from surveillance cameras connected over LTE, and for forwarding relevant incidents to the local control room. This use case benefits from MEC which provides analysis capacity close to the network edge, as well as local break-out of network traffic. This saves video transmission bandwidth, as a lot of processing can take place close to the point of capture to reduce the amount of data forwarded, and local break out allows to consume locally-produced high volume data locally without routing them through larger parts of the network. Nokia, Vodafone Hutchison Australia and SeeTec have run a MEC proof of concept (PoC) during the Informa MEC congress in Munich. As use cases, the classification / counting of people, and the tracking of object presence/absence were demonstrated. This webinar describes the architecture of the PoC, presents the demonstrated use cases and discusses the findings of the PoC.