System design

Active tampering alarm

See movie
Learn how Active Tampering Alarm helps operators when cameras are under attack.

There are two broad categories of systems for implementing intelligent video: centralized and distributed. In centralized architectures, video and other information is collected by cameras and sensors and brought to a centralized server for analysis. In distributed architectures, the edge devices (network cameras and video encoders) are ‘intelligent’ and are capable of processing the video and extracting relevant information. A further consideration is whether the system should allow for integration of applications from different vendors.

Considerations
	Reliability and system availability – minimizing the risk of system failure and associated down-time
	Scalability and flexibility – the ability to effortlessly scale the system from a few to many cameras, as well as intelligently distribute processing across the network
	Interoperability – the ability to use system components from different vendors
	Security – making sure that only authorized personnel are allowed to access the system
	Total cost of ownership (TCO) – this includes capital costs for the system components and operational expenses

DVRs and centralized intelligence

To enable central monitoring of traditional CCTV systems surveillance video can be fed directly from the analog cameras into an intelligent video-enabled DVR. The DVR will perform the video analytics (people counting or car license plate extraction for example) before taking the remaining data, digitizing, compressing, and recording it; and distributing the resulting alarms and video output to authorized operators.

In this architecture, each analog camera is connected by a coax cable to the DVR. While this approach works adequately for small installations with a limited number of cameras, it is not scalable or flexible. Each DVR comes with a specific number of inputs and adding even one additional camera entails the addition of another DVR, which is a costly proposition. Also, since DVRs are proprietary embedded devices, they cannot be easily networked, or use intelligent video applications from different vendors, and do not support general network utilities, such as for security.

Typical DVR system - centralized

Network video systems and distributed intelligence

Network video allows for completely different strategy - distributed intelligence. Distributed architectures are designed to overcome the limitations of centralized architectures by distributing the processing to different elements in the network. The most scalable, cost-effective and flexible architecture is based on ‘intelligence at the edge’, that is, processing as much of the video as possible in the network cameras, or video encoders themselves. This architecture entails the least amount of bandwidth usage since the cameras can send out data and intelligently figure out what video needs to be sent. This significantly reduces the cost and complexity of the network centric processing model, and completely eliminates the drawbacks of centralized architectures.

If cameras for example have motion detection, then rather than streaming all the video, only interesting video that has motion in it can be sent to the monitoring station for further action and analysis. The load on the infrastructure and people involved falls dramatically. For specialized video analytics, where only the data is needed and not the video, such as, people counting, or automatic number plate recognition – running the applications in the camera has a dramatic impact since the cameras can extract the required data and send just that information, with perhaps a few snapshots.

Typical intelligent video system

Furthermore, processing video at the edge – or partly at the edge – significantly reduces the cost of the servers needed to run the intelligent video applications. Servers that typically process only a few video streams when doing the entire video processing, can handle hundreds of video streams if some of the processing is done in the cameras.

By integrating intelligent edge devices with video management systems, and dividing the load between the different parts of the network, intelligent video solutions can be created that scale effortlessly, are more flexible, and cost-effective than centralized solutions.

Integrating intelligent video from different vendors

Many manufacturers of video surveillance equipment supply intelligent video applications with their products. Often these are applications that enhance camera functionality with video motion detection. Occasionally equipment manufacturers provide other, more advanced video applications with functionality such as tamper detection, and people counting.

However, building robust and commercially viable applications for video analytics requires expertise in image analysis and, sometimes, specialized knowledge in a certain application area, such as retail or transportation. For this reason, a number of software vendors have chosen to focus on supplying intelligent video applications that solve specific needs. Together with network cameras, video encoders and/or video management software systems, these intelligent video applications form complete solutions, tailored to specific market requirements.

While this creates great freedom of choice for the end user, it also calls for compatibility and easy integration between the cameras/encoders, video management software, and the intelligent video applications. In order to be commercially attractive and to optimize compatibility, devices, software, and intelligent video applications need to be built on open interfaces (APIs) and platforms. This creates flexibility for users and enables them to design intelligent video surveillance systems that fit their needs perfectly.

axisFlexSubContent