Decentralized Smart City Surveillance Data Processing Middleware for On-Premises Video Data Analysis

Brocco, Amos and Galli, Vanni (2018) Decentralized Smart City Surveillance Data Processing Middleware for On-Premises Video Data Analysis. Technical Report UNSPECIFIED

[img] Text
Surveillance_TechReport.pdf

Download (1MB)

Abstract

Surveillance cameras are nowadays pervasive elements in our cities: they act as a deterrent against illegal activities or traffic violations, they help investigating crimes, and they can provide useful traffic and people analytics. Collecting video streams from multiple sources, possibly in real-time, poses a number of ethical, legal and technical challenges. As engineers we mainly focus on the latter issue, leaving the first two to politicians and lawyers. On the one hand, we recognize that the infrastructure needed to stream data needs to be sufficiently capable, robust and scalable; on the other hand, we know that plentiful of raw processing power and a vast stor- age capacity are required to analyze multiple video streams in real time and save the resulting information. In this paper we aim attention at those technical issues and present an on-going project focused at providing real-time information by means of a flexible processing infrastruc- ture. We tackle the robustness concern through a decentralized middleware solution, and we aim at improving the scalability of the system by using machine learning techniques to extrapolate structured knowledge as soon as possible, so that only the relevant data needs to be transmitted further in the network and stored on remote servers. This approach enables surveillance oper- ators to quickly inquire the system by means of morphological or color search criteria instead of watching hundreds of video streams, then select and automatically track interesting items within different video streams.

Actions (login required)

View Item View Item