Video analytics (also known as Video content analysis, VCA, or Video Motion Detection, VMD) consists in a software that exploits images from a CCTV camera, in order to elaborate an interpretation which characterizes a situation, and allows to warn a user about it. In opposition to image analysis, video analytics is based on the observation of the evolutions of single targets in a scene, while filtering potential perturbating effects, such as camera vibrations, or day/night illumination change, or cloudy vs. sunny changes. It aims at locating and analyzing these changes in order to get an interpretation of it.
EVITECH's video analytics solutions allow to define monitoring contexts with associated sets of conditions that the user can freely define in order to filter detection of types of moves he wants to be aware of.
Examples of situations that EVITECH video analytics products can detect are the following :
- intrusion of an individual in a site, trespassing of a limit, climbing, ...
- complex move scenario (succession of situations : location, duration, direction, shape, speed, ...)
- unadapted speed (too slow/too fast), reverse direction walk/drive,
- parking, stay duration, laid object,
- counting of successive targets, classification (by size and/or shape),
- early fire start,
- oil & gas leaks.
The video analytics function is deployed behind each analyzed camera, on the network or the circuit where the image is taken and recorded. On a large CCTV system, the network rate does not allow to provide every image everywhere in permanence, and then the video analytics function is processed on camera sub-networks.
Video analytics users require an acute detection capacity AND simultaneously the lowest false alarms rates.
EVITECH's video analytics solutions are proven for detecting 100% of searched events (single targets) when they are visible "with a naked eye" on the camera image, with (e. g.) less than 0.5 false alarms per day and camera checked and confirmed on all intruder detection applications on sites installed in 2011 with this application (several hundreds of cameras).
| Next > |
|---|