Unattended Object
How it works
The Unattended Object rule generates an event when an object such as a bag, suitcase, backpack, or box is left behind in a defined region of interest, with no person remaining nearby, for longer than a set period of time. It is built for public-safety and security scenes — transit hubs, airports, stations, lobbies, and platforms — where an item left on its own may need to be investigated.
How it detects
IRIS+ detects and tracks objects and people in the scene. When an object is placed and stays in the region of interest, and no person is present near it for the configured dwell time, the rule treats it as unattended and generates an event with a short clip and a bounding box around the object. This is chained logic: the rule is not just detecting an object, it is confirming that the object has become stationary and that no person is attending it. Because of that, detection quality depends on being able to see both the object and whether people are around it.
How to set up
- Open the camera you want to configure and add a new rule.
- Select Unattended Object as the rule type.
- Draw the region of interest where left-behind objects should be detected.
- Set the dwell time — how long an object must be left unattended before the event triggers.
- Set the sensitivity (start with Normal and increase only if items are being missed).
- Save and activate the rule. It can run alongside other rules on the same camera and be scheduled to active hours.
What to expect
- An event is generated once an object has been left with no person nearby for the full dwell time.
- Every event includes an event clip and a bounding box around the object.
- Best results come from clear views of the region where objects may be left, with the surrounding area also visible so the rule can tell whether a person is attending the item.
- Busy, cluttered, or highly reflective scenes can increase false alerts; starting at Normal sensitivity and tuning the dwell time helps reduce noise.
- This rule is intended for standard camera views and is not supported on overhead / fish-eye cameras.