Video analytics, also known as Video content analysis, refers to use of technology (special algorithms) to process digital video in order to perform security-related functions.
There are several potential uses of video analytics. It can be used for automatic face recognition, vehicle number plate recognition, detect speed of vehicles in motion, count objects, detect loitering in an area, and more.
The basic idea behind video analytics is to use computer and an algorithm to detect events or instances which will interest humans, and even warn someone of a possible incident.
It is believed that video analytics can be of great help in providing security to critical infrastructures. These can help effectively detect intruders, track people as well as objects, and can be made to trigger an alarm based on certain types of behavior.
Here are a couple of examples where video analytics can be applied:
Walking past any closed store at late evening hours can be considered normal but lurking near the store may not be considered normal. It could be an indicator that the person is trying to break-in.
It is common for people to walk along the outside of a fence but if someone is trying to climb a fence, that can be a danger. So horizontally motion across the field of view is no cause for alarm but if a stranger’s action results in vertical motion across the field of view, then the video analytics algorithm can direct the video to the main people.
The bottom line is that there are just too many CCTV cameras installed all around people and it is just not practical for all the footage to be seen by just a handful of operators.
In most cases, incidents or events are discovered only after it happens. The recorded CCTV footage could provide clues about the culprits but it will take human operators a lot of time and efforts to go through all the recorded footage. This is another area where computer video analysis can be such a great help in order to dramatically cut down the search time.