AI algorithm

Off-Post and Sleeping-on-Duty Detection

AI video analytics for configured duty posts, control rooms, and workstations where staff presence, absence, or sleeping-on-duty events require review.

Request a Demo View Products

Off-post detection is intended for a defined duty location rather than general staff surveillance. A project identifies the workstation, the working period, acceptable short absences, and the evidence required by the supervisor. The system then records an event when the agreed rule is met.

Why duration matters

An empty chair for a few seconds can be part of normal work. A useful rule uses a configured duration, a schedule, and an alarm interval. This limits repeated notifications while allowing the supervisor to see the relevant time and image. Shift changes and meal breaks should be excluded before the system goes live.

Limits of posture recognition

Possible sleeping posture can be a useful prompt for review, but it should not be treated as a final personnel decision without the full camera context. Low light, face masks, desk partitions, and a person leaning down to work can affect the result. A camera angle that clearly shows the upper body is more useful than a broad room overview.

Deployment record

For each duty point, keep the camera name, working schedule, detection area, absence duration, review owner, and exception process. This makes later adjustment straightforward when the post arrangement changes.

Algorithm deployment notes

Off-post and sleeping-on-duty events need a site-defined review process

Duty-room and post-presence analytics examine visible occupancy and activity patterns in a defined camera scene. The operational rule needs to be agreed carefully because a person temporarily out of view, reading a screen, resting during an authorised break, or obscured by equipment may not indicate the same condition.

Choose a stable duty-post view

The camera should show the assigned position, desk or work area, normal access path, and the part of the scene needed to assess presence. Avoid views where glare, monitors, partitions, frequent visitors, or an extreme side angle hide the person or make posture difficult to review.

Use schedules and authorised breaks

Configure the monitored area, duty schedule, minimum absence or inactivity period, alarm repeat interval, and any authorised-break exception. This allows the site to distinguish an operationally relevant absence from a normal handover, break, or approved task outside the desk.

Route events for discreet review

Event evidence should be limited to authorised supervisors and include the time, camera, configured rule, and reviewable image or video reference. The workflow should respect privacy and employment requirements while allowing the responsible manager to verify a possible issue.

Validate the wording and response

Before wider use, review sample events with the duty management team and agree what counts as a follow-up, a false alarm, or an authorised exception. The value comes from a clear response process, not from treating every automated observation as a final judgement.

Related planning: AI NVR recording and event management. Access rights, retention, and review procedures should be defined before such analytics are enabled.

Project inquiry

Tell us your camera environment and required AI analytics.

For the first version this site uses a lightweight email-based inquiry flow. It avoids a database and keeps the English site simple until inquiry volume justifies a dedicated CRM integration.