Big Data Problem
In London, over half a million cameras observe public space; that’s one camera for every 16 people. Assuming that video is stored for one week, the video storage is at least 100 petabytes. Locating a subject, a vehicle or an event in this massive video recordings is like looking for a needle in a haystack.
A new breed of video surveillance systems is emerging, which are optimized for searching big video data, like Google searches the Internet.
One of the examples is Kipod developed by a startup in Toronto. Kipod leverages a cutting edge cloud architecture and machine learning technology to provide a next generation experience in large scale CCTV installations.
What is Kipod?
Kipod is the industry’s first carrier class cloud platform for video content analytics and alarm monitoring. Kipod enables real-time crime detection and instant search across big data of CCTV video using the most advanced machine learning algorithms. With Kipod, telecom operators and security providers can deliver a next generation video surveillance service to governments and enterprises, serving millions of users and millions of cameras.
Kipod users can locate an event or track subjects with no preconfigured rules set. All object trajectories and features from live video are recorded as metadata and are made available for searching.
Kipod has a built-in engine for behavior recognition, face recognition, number plate reading and audio recognition. All search queries can be executed across thousands of computers within a private cloud using a web browser or mobile app. No bulky desktop application is required.
What is different in Kipod?
Unlike traditional video management systems, the Kipod cloud platform provides important carrier class features:
- instant video content search across big data from any web browser and mobile device
- scalability beyond 1M users, 1M cameras and 1M petabyte (1 exabyte) video storage
- tier 3 availability with full redundancy and background software updates
- open source software infrastructure (Linux, Kubernetes, Cassandra and Ceph)
Browsing a Map and Site Plans with Kipod
Evolution of CCTV Control Room
- 1st generation – local control room in each location. Very expensive staffing, difficult to ensure monitoring standards (response time, protocols etc).
- 2nd generation – centralized control room for multiple locations. Much more efficient, but cannot be scaled to serve large cities and large number of users.
- 3rd generation – virtual control room for multiple locations. Authorized users from local government, police, intelligence agencies can connect from anywhere to monitor locations of interest or search for events in video. Kipod fits this category by providing access to searchable video to thousands of users.
Applications of Kipod
- Urban CCTV (safe cities, public spaces, transportation)
- Distributed infrastructure monitoring (oil and gas, energy, utilities, logistics)
To learn more, please visit Kipod.com.
*Videos courtesy of Kipod and YouTube