TopTechnical DictionaryIVS - Intelligent Video Surveillance

IVS - Intelligent Video Surveillance

Intelligent Video Surveillance (Intelligent Video Surveillance) is a video processing technique detecting events based on pre-defined criteria.

 

Depending on the model, the following IVS functions are available:

 

Tripwire (Tripwire) – a function tripping an alarm after a line drawn is intersected by an object moving in a specific direction. A direction of the object movement in relation to which the alarm is tripped must be configured. Available configurations: A->B, B->A or movement in both directions A<->B.

 

The following video shows the Tripwire (Tripwire) function:

 

 

 

 

Intrusion (Intrusion) – a function detecting an object entering or leaving a secure zone and tripping a pre-defined alarm.

 

The following video shows the Intrusion (Intrusion) function:

 

 

 

Abandoned/Missing (Abandoned/Missing) – a function detecting an abandoned object in a secure zone or indicating an object missing from the secure zone.

 

Fig. 1. Screenshot of the Abandoned/Missing (Abandoned/Missing) function

 

The following video shows the Abandoned (Abandoned) function:

 

 

 

The following video shows the Missing (Missing) function

 

 

 

Scene Changing (Scene Changing) – a function detecting any change in the image. The device uses an intelligent scene comparison technique. If the comparison shows that the scene has changed, a pre-defined alarm event is triggered. The function indicates if the camera is obscured (see video below - scene 1), if its position has changed (see video below - scene 2) or if any other irregularity is detected.

 

 

 

People Counting (People counting) – a function showing statistical data for objects entering and leaving a secure zone. Daily, monthly and annual reports are available. The tool generates linear or bar graphs and all the data can be exported.

 

Function performance can be affected by:

 

1. Camera installation angle – an angle between the camera and the vertical line should be 40° to 60°.

 

2. Person’s head – the head size should be at least 100x100 px in 1920x1080 px images. Recommended size: head takes 1/18th to 1/7th of the frame.

 

3. Background – set the camera in a location with a uniform background.

 

Fig. 2. Uniform background (simple background) example

 

The following video shows the People Counting (People Counting) function

 

 

Fig. 3. Screenshot of the bar graph

 

Fig. 4. Screenshot of the line graph

 

Face Detect (Face Detect) – a function detecting faces within a frame. Face detection may trigger a photo, a video or an alarm event.

 

Performance of the face recognition algorithm can be affected by: 


1. Angle between the camera and the horizon – recommended angle from 0° to 30°.

2. Face size and position - angle of face inclination towards the camera, face visible in the frame must be at least 100x100 px.

3. Illumination – uniform and without reflections.

 

Fig. 5. Screenshot of the Face Detect (Face Detect) function

 

The following video shows the Face Detect (Face Detect) function

 

 

 

Heat Map (Heat Map) – a function analysing the movement patterns. The function allows to analyse the population density in time. The systems can be use in the shopping centres to measure the traffic intensity at a specific date and time, to estimate the number of customers and highlight the places of interest. The heat map shows the traffic intensity using different colours. Red indicates high density, blue indicates low intensity.

 

Fig. 6. Screenshot of the Heat Map (Heat Map) function

 

Fig. 7. Screenshot of the Heat Map (Heat Map) function

 

Audio Detect (Audio Detect) – a function evaluating irregularities at the audio input. Detected irregularity may trigger a picture, a video or an alarm event.