The first video analytics, used in analog systems, were centralized. In these installations, video was transferred to the Digital Video Recorder (DVR), where the analysis took place. This meant that all video had to be transferred, often with many hours of video containing no content of interest, all of which generate excessive network loads and storage requirements. This process also required additional and costly servers, to process these huge amounts of data.
The latest generation of video analytics –distributed or edge analytics – distributes the video processing to the location where it makes most sense, i.e. in the camera or the video encoder. Analyzing at the edge means that no dedicated analytics servers are needed, and that analysis is performed on the uncompressed video feed, creating a much more cost-effective and flexible architecture.
Figure 3: A truly distributed system with intelligence at the edge (in the network cameras and video encoders) provides the most scalable and cost-efficient system.