Detection Mode
Every camera CCU runs one detection pipeline at a time, and Detection mode is where you choose which one. Almost every hub stays on Normal, the intrusion pipeline that has always run. Pose is a specialist mode for sites that need to know about people's posture rather than their presence — a person putting their hands up, or a person lying on the ground.
The mode belongs to the hub, not to individual cameras. When you change it, the hub restarts its detection service and every camera on that hub switches together.
Normal mode
The default. The hub watches for people and vehicles, verifies detections against the cloud verifier, and raises intrusion events according to each camera's own detection settings. Nothing about this behaviour changes — a hub left on Normal behaves exactly as it always has.
Pose mode
Pose mode replaces the intrusion pipeline entirely. Instead of asking "is there a person here?" the hub asks "what is this person doing?" and raises an event when it sees one of two postures:
- Hands Raised — both wrists held above the shoulders.
- Person Lying Down — a person horizontal on the ground.
This suits sites where people are present all day and their mere presence means nothing — a hold-up at a counter, or a person collapsing in an area that is normally busy.
What pose mode changes
| Normal | Pose | |
|---|---|---|
| Detects | People and vehicles | Hands raised, person lying down |
| Cameras per hub | Up to 10 | 5 |
| Camera codec | H.264 or H.265 | H.264 only |
| Resolution per camera | Up to ~1 MP | Up to ~1080p |
| Cloud verification | Yes | No — events go straight through |
Three of those need explaining:
Five cameras. Pose detection runs a second AI pass on every person in every frame, which costs far more processing than intrusion detection. Five cameras is what a hub can sustain. If you select Pose on a hub that already has more than five cameras, CleverOps blocks the change and tells you how many to remove first. Cameras beyond the fifth on a hub already in pose mode show as Over Budget and are not analysed.
H.264 only. Pose mode refuses H.265 (HEVC) camera streams. H.265 forces the hub to decode in software, which would consume the capacity that pose detection needs. A camera streaming H.265 on a pose hub shows the status "Blocked: H.265 not supported in Pose mode" — change the codec in the camera's own settings to H.264 and the hub picks it up again within five minutes.
Higher resolution allowed. Because pose mode runs fewer cameras and rejects the expensive codec, each camera is allowed a larger picture — up to about 1080p, versus roughly 1 MP in Normal mode. Pose detection reads shoulder, wrist and hip positions, so detail on distant people matters much more than it does for intrusion.
How pose events are handled
Pose events skip the cloud verifier. That verifier is trained to confirm intruders in intrusion footage and has nothing useful to say about posture, so a Hands Raised or Person Lying Down event goes straight to the control room as raised.
To avoid alarming on a single misread frame, the hub requires three sightings of the same posture within 30 seconds before it raises an event, then stays quiet on that camera for 30 seconds afterwards.
Events arrive in the control room named Hands Raised or Person Lying Down.
Changing a hub's detection mode
- Open the site the hub is installed at, and go to the Hardware tab.
- Find the hub in the list and click Edit on the hub's own row (not on one of its cameras).
- Set Hub detection mode to Normal or Pose.
- Click Save hub settings.
The camera editor has its own Detection mode setting under Camera onboard analytics — that one controls whether an individual camera's built-in analytics are used alongside the hub AI. It appears only on cameras connected to a CCU; CleverMail cameras show their Camera alarm types instead. Hub detection mode is the hub-wide setting described on this page.
The hub restarts its detection service and comes back in the new mode within about a minute. Cameras are offline for detection during that restart.
Detection mode requires CCU firmware 1.13.10 or later. Older hubs ignore the setting and stay on the intrusion pipeline, so check the hub's version before switching a site over.
If the hub has more than five cameras, remove or disable the extras before switching to Pose — CleverOps will not let you save until the hub is within the limit. Switching back to Normal restores the full camera allowance, but you will need to re-enable any cameras you disabled.
Troubleshooting
A camera says "Blocked: H.265 not supported in Pose mode". The camera is streaming H.265. Change it to H.264 in the camera's own web interface or through the camera setup wizard. The hub re-checks every five minutes.
A camera says "Over Budget". The hub is in pose mode and this camera is beyond the five-camera limit. Disable a camera you care less about, and this one starts analysing.
The mode did not change. The hub must be online to receive the switch. If it was offline when you saved, it applies the change automatically the next time it comes online and syncs its cameras.