In security and surveillance systems, a major task has been finding an affordable way to move, store and later search through hours, days, weeks or even months of video images. At this point, the surveillance industry is going through a major upheaval due to key technological developments.
These developments include the emergence of affordable digital cameras with low-cost mega-pixel capabilities; low-cost high-definition (HD) monitors for display; new digital signal processors (DSPs) with compression co-processors; the pervasiveness of Internet Protocol (IP) technology and convergence to a single site network; cost savings of power over Ethernet (PoE) so electricians are not required to install power per camera; and the shrinking cost of mass storage. Further, next-generation cameras have emerged with built-in computing power that allows for the implementation of video analytics, which offloads processing tasks and otherwise assists overburdened human operators.
These are the essential trends helping security personnel handle, display, evaluate and store the large amount of video collected from the increasing number of surveillance cameras that are being deployed everywhere.
Bandwidth, Latency, Storage Remain Major Challenges
Security practitioners are well aware that the video requirements for surveillance are far different from “Hollywood” video. These are not studio conditions. Often cameras are situated in environments with little light and contrast, and camera placement is not always ideal for capturing every location of interest.
Luckily in this regard, security systems do not generally need broadcast market image quality if the system is mainly used to monitor for people, objects or motion in a given area. For higher end applications, the system may need a high enough resolution to capture and read license plates in poor conditions when a vehicle is moving.
North American television has a frame rate of 30 frames/second (FPS) and an active image resolution of 720 X 480 pixels. Security systems can tolerate far lower resolutions and frame rates because every minute detail does not need to be captured. This allows for a considerable tradeoff in terms of storage and channel density. For these reasons, many analog-based security systems run at 7.5 FPS and/or an image resolution of 352 X 288 (CIF).
Reducing either the frame rate or resolution allows four camera images in the bandwidth, which is normally taken up by one. Reducing both the frame rate and resolution allows 16 camera images in the bandwidth normally taken up by one. Of course, saving this bandwidth carries a cost — the image quality goes down. While motion may be jerky at 7.5 FPS and 352 X 288 images are not as detailed, the operator can still generally determine what is occurring within the scene. This is the situation with legacy systems, where one or both of these image quality concessions are made to keep costs down.
New installations are utilizing improvements in compression and cheaper storage so security systems can process video at higher frame rates and resolutions — often a full 720 X 480 pixels and 30 FPS. If needed, these systems can still be programmed to process video at lower frame rates or resolutions to help reduce costs.
Another key requirement is low latency, both in terms of how long it takes video to appear on a display as well as in how quickly the camera can pan and zoom to a suspicious event. Live video surveillance can sometimes be hampered when latency is so high that joystick control is out of sync with camera movement. Analog TV systems have historically performed better compared to IP video systems in terms of transmission latency.
The challenges with low-latency digital video include overcoming the highly variable delays and losses of Ethernet networks and the inherent latency added by the video coding and decoding. These challenges are typically met with a good network infrastructure as well as employing advanced codecs and video-compression techniques.
Video storage is also a challenge since a majority of security systems have to be able to store and later recall video frames from days, weeks, months or even years prior. A system with a combination of efficient compression algorithms and low-cost mass storage can hold the necessary amounts of data, but given this large amount of data, it also needs to be able to mark and jump to specific sections, such as when people enter into a frame or when a particular car leaves a parking lot.
Page 1 of 3 pages 1
Video Surveillance ·
Digital Signal Processing ·
Image Quality ·
Power over Ethernet ·