Human Detection
Identifies people in the field of view to enable person-based rules and alerts. It is where most sites start, because separating people from everything else is what stops a motion rule alarming on rain.
What triggers it
A person enters the frame and stays long enough to be tracked. Reflections, posters and screens showing people are the common source of false positives, so it discriminates on movement and depth rather than shape alone.
What you set up
- the zone
- minimum time in frame before it counts
- confidence floor
What the operator gets
A bounding box on each person with a confidence figure, and a count of people currently tracked in the zone.
Where it’s used
Every site defines this differently.
What counts as an event in a hospital lobby isn't what counts on a station platform or a factory floor. Describe your scene and we'll tune the detection to it.