Video analytics for security, able to autonomously detect situations in the field of view of the cameras, can be used to react (reactive algorithms) or to monitor (monitoring algorithms). They can also be functionally classified by categories of monitored and / or observed objects. The main categories are humans, vehicles, animals, smoke and fire, liquids, food and also user-definable objects (e.g. trolley, backpack, suitcase, etc.).
Some examples are:
- The detection / recognition of faces [Face Detection / Face Recognition] and the relative comparison with the faces of people present in an archive in addition to the estimate of the ethnic group of the detected person.
- Thermo-scanning [Fever Detection] and the detection of the presence of a person’s mask is carried out for the purpose of preventing any epidemic / pandemic contagion.
- The recognition of a license number plate [LPR / ANPR] of a vehicle, which can be extended in some cases to the brand, the model and the color of the vehicle.
- Crowd detection [Crowd Detection] Given a minimum number of components that can be defined by the user, it allows you to recognize a group of people (crowd) that exceeds this minimum value and indicates its number.
- Object detection [Object Detection] detects people, pets, vehicles (even without license plates), suitcases (bags, trolleys, backpacks, etc.) and other user-definable categories of objects.
- Smoke and Fire detection [Smoke and Fire Detection] detects the onset of fire.
- Motion detectors [Motion Detection] and behavioral analysis [Behavior Analysis] are algorithms that are activated on the basis of variations in the visual field [Field of View] such as eg. line crossing, loitering, tripwire, abandoned object, etc.
- The combined analysis of vehicular and human recognition based on AI allows operations such as the management of an unattended automatic gate [Parking Automation].