Situational video analytics

 

 

Situational video analytics allows us not only to detect an object and track its movements, but also to classify the object’s behavior on the basis of parameters set by the user. Situational video analytics considerably decreases the workload on an operator because it removes the necessity of analyzing all the objects within the camera’s view thus excluding the influence of the human factor on the work of the system.
Signal line: crossing the line, crossing the line in pairs (movement at the rear in close proximity), crossing without a pass/ticket, movement against the stream, a great number of people entering the territory simultaneously.
Signal zone: entering the zone, leaving the zone, stopping, purposeless movement, fast movement (run), movement in a certain direction, crowding of people, left/lost object.

 

 

Face detection

 

 

The facial detector helps detect and track all the persons within the camera view. All persons are stored in an archive which enables us to find anyone who has ever visited the object, for example to find people for a certain period of time or those who entered the object through a certain entry. The algorithm is optimized for cameras of very high resolution used in existing facial detection systems.

 

Number plate reading

 

 

The number plate reading (NPR, or license plate reading – LPR) module is a technology that has become crucial for today’s security requirements.
The number plate reading module has a high quality, is not demanding to hardware resources, and is easy to operate.

 

Crowd detector

 

 

The crowd detector is designed to prevent crime and civil disorder in the territory under surveillance. The detector may be used both in the streets and squares and inside buildings: metro, stations, trading and entertainment centers.

 

Fight detector

 

 

The fight detector automatically detects the unusual behavior of people in terms of disorderly conduct, robbery, fighting. There are no similar solutions in the video surveillance market today. In case of algorithmic video surveillance, the hardware and software complex does the initial detection, but it is the operator who makes final decision. Thus the module helps control almost a hundred video streams without missing any key events.

 

Left objects

 

 

The detector of left (or lost, not belonging to anyone) objects is designed to determine situations when an object left in the surveillance zone, for example an explosive, may be potentially dangerous. To find an object left in a crowd of people is a complicated scientific and technical task, which is still not resolved completely by any producer of video analytic systems.

 

Video quality control

 

 

The tampering alarm enables us to continuously monitor the quality of a video signal, to detect any facts of unauthorized interference with the video surveillance system or any unexpected violations of the surveillance conditions. It may be, for example, the camera being turned in a wrong way, the installation of mirrors that may cause disorientation, the lost signal, blackout, lighting up, or the loss of focus. This function is used for both outer and inner surveillance cameras.