Loitering detection is video analytics that raises an alert when a person stays in a defined area longer than a set time, so a security team hears about it while the person is still there. Instead of a guard watching a wall of screens for someone who lingers near a loading dock, a gate or a fence line, a software agent watches every feed and reports only the cases that match the rule. Argu’s intrusion detection agents apply this kind of rule to cameras a site already has, without new hardware.
Key takeaways
- A loitering rule has three parts: a zone, a dwell time and a condition, such as a person present after hours.
- The dwell time is the setting that decides how many alerts a team receives, so it should be tuned per zone and not set once for the whole site.
- Loitering is a signal, not proof of intent. Alerts should go to a person who checks the video before anyone acts.
- Argu reads existing RTSP or ONVIF camera streams at 1080p to 4K and 15 to 30 frames per second, and a new rule written in plain English goes live in about 30 seconds.
- Detection events and the matching video can be kept for an operator-set window of 30 to 90 days.
What Is Loitering Detection?
Loitering detection watches a camera view for a person who remains in one area for longer than expected. The system does not need to know why the person is there. It measures how long they stay and compares that with the rule the security team wrote. If the time limit is passed, it sends an alert with a clip of the event.
The idea is simple, but the value is in what it replaces. A guard cannot stare at forty feeds at once, and a motion sensor cannot tell a person who walks through a lot from one who waits beside a gate for twenty minutes. Dwell time is the missing piece, and it is the reason loitering rules catch behavior that motion alarms ignore.
How is it different from motion detection?
Motion detection reacts to pixels changing. Wind, headlights, rain and shadows all change pixels, so a motion alarm cannot separate a branch from a person. A loitering rule starts with a person, then tracks that person across the zone and counts time. For a closer look at why raw motion alarms create noise, see our guide to false alarm rate.
How is it different from intrusion detection?
Intrusion detection asks whether someone crossed a line or entered a zone that is closed to them. Loitering asks how long someone stayed in a zone that may be open to the public. A visitor walking through a parking lot is not an intrusion, but the same visitor standing beside a locked service door for fifteen minutes may be worth a look. Our article on perimeter intrusion detection covers the fence-line side of the same problem.
How Does a Loitering Rule Work?
Every loitering rule combines the same building blocks. Getting each one right is what separates a useful alert from one that gets ignored by the second week.
1. The zone
The zone is the part of the camera view where time is counted. A narrow zone around a door or a cash machine produces cleaner results than the whole frame, because people walking past the edge of the scene are never counted.
2. The dwell time
The dwell time is how long a person may stay before an alert fires. Short limits catch more events and produce more false alarms. Long limits are quiet but may fire after the moment has passed. The right value depends on the place, so start with a number your supervisors agree is unusual and adjust after the first week of alerts.
3. The condition
A condition limits when the rule is active. The most common ones are time of day, such as outside business hours, and context, such as a person near a vehicle or a door. A rule that only applies at night will stay silent while a lobby is busy during the day.
4. The response
An alert is only useful if it reaches someone who can act. The response step decides who receives the clip, how fast, and what they are expected to do. Argu agents send the alert with the video, so the person on duty sees what triggered it before they pick up a radio.
Where Is Loitering Detection Used?
The same rule fits many places, but the zone and the dwell time change. The table below shows example settings that a team might start with. They are starting points for tuning, not recommended values.
| Location | Example zone | Example starting dwell time | Example condition |
|---|---|---|---|
| Warehouse loading dock | Area beside the dock doors | 5 minutes | Outside shift hours |
| Retail store exit | Space just inside the doors | 10 minutes | After closing |
| Parking lot | Rows beside parked vehicles | 3 minutes | Between 10 pm and 6 am |
| Utility fence line | Strip along the outside of the fence | 2 minutes | Always on |
| Port gate | Pedestrian side of the gate | 5 minutes | Person without a vehicle |
Warehouses and yards
Loading docks and yards often have long quiet periods and a few busy ones. A loitering rule tied to off-hours gives a supervisor a reason to look at a dock at 2 am, which is the kind of check a guard rarely has time for. Our guide to warehouse surveillance shows how the same cameras serve several rules at once.
Parking lots and garages
A lot has many people with a good reason to be there, so the rule should focus on what is unusual: a person lingering between vehicles late at night, for example. Pair it with plate reading when the question is also which vehicle was there, as described in our guide to parking lot surveillance.
Critical infrastructure and ports
For utilities, ports and other regulated sites, a person standing beside a fence for several minutes is an early warning that is worth acting on before any climbing starts. This is one of the reasons Argu’s perimeter agents look at distance to the fence and behavior, not only at a crossing line. Read more in our overview of critical infrastructure security.
Retail and commercial venues
In stores, loitering rules work best at specific points such as an emergency exit or a stockroom door, where lingering is unusual by design. They should not be used to follow ordinary shoppers who browse for a long time. Where a retail team needs counts and queue data, people counting is the better tool.
How Do You Reduce False Alarms?
Loitering alerts fail for one reason more than any other: too many of them are about people who have a good reason to be there. A short checklist keeps the signal strong.
- Draw the zone around the sensitive spot, not the whole view.
- Set a dwell time that is longer than a normal pause, such as waiting for a phone call or a lift.
- Add a time-of-day condition so the rule is quiet when the place is busy by design.
- Review the first week of alerts with the people who receive them and move the dwell time up or down.
- Remove or reshape zones that produce repeated alerts for staff routines, such as a smoking area.
Argu agents are described in plain language, so a supervisor can change a rule without waiting for an engineer. If an alert keeps firing on a routine, rewrite the sentence and the new rule is live in about 30 seconds.
What Are the Limits of Loitering Detection?
Loitering detection tells a team that someone stayed. It does not tell them why. A person waiting for a ride, a courier looking for a delivery door and a person casing a building can look the same on camera. The alert should lead to a human review of the clip, and any action should follow the site’s own policy.
Camera placement matters too. If the zone is too far from the lens or hidden behind a pillar, the system cannot track the person. Privacy rules also apply. Retention should be set by the operator’s own policy, and Argu lets a site choose a window of 30 to 90 days for stored events. Check local law on video recording and signage before turning on any rule that watches public or semi-public areas.
Frequently Asked Questions
What is loitering detection?
It is a video analytics rule that alerts a team when a person stays in a defined area longer than a set time. It measures dwell time and does not judge intent.
How long should the loitering time limit be?
It depends on the place. A fence line or a locked door can use a few minutes, while a busy lobby needs a longer limit. Start with a value your supervisors consider unusual and adjust after reviewing a week of alerts.
Can loitering detection work on cameras I already have?
Usually yes. Argu reads existing RTSP or ONVIF streams at 1080p to 4K and 15 to 30 frames per second, so most modern IP cameras can be used without replacement.
Does loitering detection identify who the person is?
No. A loitering rule counts time in a zone. Identifying a person is a separate capability and should only be turned on where the law and the site policy allow it.
How is loitering detection different from motion detection?
Motion detection reacts to any change in the image, including wind and lights. Loitering detection starts from a person in a zone and measures how long that person stays.
How long are loitering alerts and video kept?
The operator sets the window. Argu supports a retention period of 30 to 90 days for detection events and their video, based on the site’s own policy.
See a loitering rule on your own cameras
Tell us which door, dock or fence line worries you most, and we will show how a rule for it would be written and tuned. Argu AI
Last updated: November 2026
