Computerized Video surveillance
Video surveillance is not a new concept. The use of video cameras to watch over an area didn’t need much imagination or invention; it was always obvious. However, the very first CCTV (Closed Circuit Television) cameras that were used for video surveillance were very primitive. They were big, easily visible, could not zoom in on objects, or pan around to see a more complete view of the area. Moreover, they generated low-quality black and white video, which could not be used to detect whether objects were people, chairs or curtains, much less recognize faces or colors.
Today’s video surveillance, on the other hand, is much more advanced. It consists of video cameras that are so small that they cannot be easily spotted. In addition, these cameras can pan around, zoom into an object, and record video in full color and high definition. Thus, the minutest details can be observed using the new generation of CCTV cameras. But, the older generation of poor-quality cameras has ensured that any new cameras are often said to have been ‘oversold’. “Oversold’ simply means that so many institutions and organizations have spent so much of their funds on installing the older, poor-quality cameras, that people generally think of video surveillance as “poor quality surveillance”.
Another problem for video surveillance is that, with the older cameras, many features have also been ‘oversold’, which did not live up to expectations. For example, many of the older systems were marketed as being able to ‘detect unattended luggage’. The older systems, however, were not intelligent enough to understand whether an object was ‘luggage’, and further, if the luggage was actually ‘unattended’. This further worsened the reputation of video surveillance systems.
The truth couldn’t be any more different for video surveillance. New systems have incorporated so many advanced features that the old systems seem like ‘cavemen cameras’. The key technology in this revolution is known generically as “Video Content Analysis” (VCA), though it may be implemented differently by different vendors. VCA, as it is popularly known, works by analyzing the video generated by the cameras. After analysis of moving objects, it can differentiate between walking, crawling and moving cars. Some new video surveillance systems are so advanced in VCA that they can check people’s faces and determine their approximate ages.
VCA implementations in video surveillance can also identify the location, pattern of movement, and any variance in the movement of a person. Using the computers linked to the cameras, users can also specify rules that help them to prevent or detect events, such as “raise an alarm when any person approaches the safe door”, or “Alert security silently if someone stays too much time in an ATM without using it”. Using such rules, the video surveillance system becomes more than a bunch of cameras, and can recognize unusual behavior, such as a person lingering for too much time in a parking area, which is suspicious.
Video surveillance with VCA can even identify people, point out their positions on a map, and generally make life easier for security guards worldwide.