Technology

Video Analytics Architectures

 
Video surveillance systems typically include the following main components:
  • Video cameras
  • Network infrastructure
  • Security management solutions (Video Management Systems, Command & Control Systems etc.)
  • Storage
  • Video Analytics
Video Analytics can be implemented in three different configurations, which correlate to the evolution of the Video Analytics and surveillance technologies:
 
1.       Server Based Implementation
In this approach, the Video Analytics is implemented through a dedicated server that pulls the video, analyzes it, and issues the alerts or analysis results. This approach is independent of the video cameras, and therefore, is applicable to most types of surveillance systems. The main disadvantages to this approach are:
  • The Video Analytics server requires the video to be transmitted to such server, and therefore causes an increase in network traffic load;
  • The video quality being analyzed by the Video Analytics server is usually degraded due to compression and transmission effects, and therefore, the Video Analytics performance may be compromised;
  • The Video Analytics server is limited by its processing power, and can typically handle no more than 16 cameras, with only limited Video Analytics functions, which makes it unattractive to large scale surveillance installations which deploy dozens or hundreds of cameras requiring a variety of Video Analytics functionalities.
2.       Edge Based Implementation
In this approach, the Video Analytics is implemented through an IP video camera or video encoder, which must have sufficient processing power to run the Video Analytics functionality. On the surface, this approach seems ideal, however it does not perform satisfactorily in many cases as it imposes limitations on the overall surveillance system design and performance. Most edge devices still lack sufficient processing power for high-end Video Analytics requirements, and therefore such implementation compromises on either the range of functions or performance quality of the Video Analytics, or both. In addition, most surveillance installations include different types of cameras, and not all cameras are suitable for “edge based implementation” nor do all cameras support it to the same quality.
3.       Agent Vi’s Distributed Implementation
Agent Vi has a unique and patented architectural approach to Video Analytics, called “Image Processing over IP networks” (IPoIPTM). With Agent Vi’s architecture, the Video Analytics task is distributed between the edge device (which may be an IP camera or encoder) and a server. This approach optimizes the workload on the edge device and server and yields high quality analytics performance. A key benefit to Agent Vi’s distributed architecture is that a single server can run comprehensive Video Analytics functions on up to 200 cameras simultaneously. This hardware efficient camera-to-server ratio is achieved without compromising on the range and performance of the analytics functionality, which makes it especially beneficial for large scale surveillance installations. Read more about Agent Vi’s IPoIPTM distributed architecture and competitive advantages.
View a table comparing the advantages and disadvantages of different video analytics architectures.

image description

View Vi-System in Action – Detecting a suspicious object, great for major transportation hubs and other sensitive sites

Download
The Milestone-Agent Vi partnership underlines the valuable role that analytics can play in the surveillance world, where companies who glean more intelligence from their solutions will reap the best return on their investment. Henrik Friborg , VP Strategic Alliances and Co-Founder of Milestone Systems

Contact Agent Vi