Here are several sample pipeline configurations that solve typical object recognition tasks. Get one of the available templates and load it while creating a new template.

create pipeline

Vehicle counters

This vehicle counting pipeline detects vehicles, tracks their movements and, if a vehicle crosses virtual line, generates a 'line_cross' event, that can be sent to an Output.

Pipeline element Description
Vehicle detector The neural net searches for all vehicles in the frame
↓
General object tracker The tracking processor builds a track for each detected object, adding information about objects' movements.
↓
various filters Add filters here to remove unwanted objects
↓
Line cross detector The line cross processors checks if an object's path has crossed the virtual line and adds a 'line_cross' event.
↓
System DB System DB output saves all line_cross events to the local DB.
↓
Notifications In addition to the statistics, the notification output allows sending warnings if someone crosses a line.

Downloads:

License plate recognition

This pipeline detects and recognizes license plates. It generates 'lp_detected' event, that can be sent to an Output.

Pipeline element Description
License plate detector The neural net searches for all license plates in the detection area
↓
General object tracker The tracking processor builds a track for each detected object, adding information about objects' movements.
↓
License plate recognition The recognition converts license plate image into symbols.
↓
General event trigger This trigger is needed to generate 'lp_detected' events.
↓
System DB System DB output saves all events to the local DB.

Downloads:

People gender/recognition plus counter

This people counting pipeline detects faces, recognizes age/gender, tracks their movements and, if a person crosses virtual line, saves this event to the DB.

Pipeline element Description
Face detector The neural net searches for all faces in the frame
↓
General object tracker The tracking processor builds a track for each detected object, adding information about objects' movements.
↓
various filters Add filters here to remove unwanted objects
↓
Age, gender, emotions recognition The neural net recognizes person's age/gender.
↓
Pixelate This processor can apply a pixel mask to the faces to match privacy requirements.
↓
Line cross detector The line cross processors checks if an object's path has crossed the virtual line and adds a 'line_cross' event.
↓
System DB System DB output saves all line_cross events to the local DB.

Downloads:

Important

After you create the pipeline, don't forget to update 'active area', 'counting lines', etc and add an Output to the bottom of the pipeline (for example a [SYSTEM_DB])