Ayonix Video Analytics

Industry

Video analytics for manufacturing plants and production sites

In short

Manufacturing video analytics turns plant cameras into schedule-aware events: presence in areas that should be empty, gate and yard movement, and area occupancy. Rules are built around shift patterns so routine production activity generates no noise and genuine exceptions reach a supervisor.

Wide view of workers, forklifts, conveyors and cameras in a distribution centre
Conceptual illustration of distribution-centre operations. AI-generated conceptual image, not a customer deployment.

The problem

What factories teams are dealing with

  • Restricted areas are entered outside permit windows and it surfaces only in an audit months later.
  • Motion-based rules in plant environments fire constantly on steam, moving machinery and forklift lights.
  • Gate and yard movement is recorded but there is no searchable record of who arrived when.
  • Shift patterns mean the same camera needs completely different rules at different times and no system expresses that.
  • Plant cameras were installed for general surveillance and nobody has assessed whether they can support a rule at all.

Camera placement

Where cameras need to be, and why

Camera position decides more of the outcome than any software setting. These are the placements that matter in this sector.

Recommended camera placement forFactories
Location Purpose What matters
Restricted process areas Schedule-aware presence rules Exclusion sub-zones for permanently staffed positions are essential, or the rule fires continuously.
Site perimeter and fence line Intrusion detection Fixed, stable mounting; plant vibration and wind movement are a leading nuisance-alarm source.
Goods gate and weighbridge Vehicle detection and crossing Views across the lane rather than along it, positioned away from habitual queueing.
Yard and trailer bays Vehicle presence and dwell Elevated placement so the front row does not hide the rows behind it.
Plant room and substation access Out-of-permit presence IR performance at the required range must be verified, not assumed.

Workflow

From event to acknowledged action

Event workflow

  1. 1 Plant camera streams are ingested and monitored for availability and blocked views.
  2. 2 People and vehicles are detected and classified; steam, dust and machinery movement are filtered by class and size.
  3. 3 Zone, line and dwell rules are evaluated against the shift and permit schedule.
  4. 4 Events are stabilised with persistence windows before any alert is created.
  5. 5 Evidence is assembled and routed to the VMS or the supervisor channel.
  6. 6 Events are recorded in full for the safety and compliance review even when not notified.

Operator workflow

  1. 1 Shift supervisor receives an out-of-permit presence event with a clip.
  2. 2 The plant camera is opened directly from the alert and the situation assessed.
  3. 3 The response follows the plant’s existing permit and safety procedure.
  4. 4 The event and its outcome are recorded for the weekly review.
  5. 5 Recurring patterns drive a procedure change rather than more alerts.

Operations

Dashboards, integration and deployment

Dashboard metrics

  • Out-of-permit entries per zone per shift
  • Perimeter events per night and acknowledgement time
  • Gate and yard vehicle movements by interval
  • Area occupancy against configured limits
  • Irrelevant alerts per camera-hour by cause
  • Camera availability across the plant estate

Integration options

  • VMS alarm and custom-event delivery
  • Webhook delivery into plant operational and permit systems
  • REST retrieval for compliance reporting
  • ONVIF and RTSP ingestion from the plant camera estate
Compatibility states →

Deployment options

  • On-premise processing inside the plant network, which is the normal choice
  • Edge processing for remote plant rooms and substations
  • Air-gapped deployment where the OT network is isolated from corporate IT
  • Multi-site management so a validated rule set is copied across comparable plants
Compare architectures →

Evidence

Deployment pattern

Plant compliance pattern

Schedule-aware restricted-zone rules, perimeter detection and gate vehicle movement, recorded in full and notified on exceptions only.

Nuisance-alarm reduction and entry recall are measured against your own plant baseline during the pilot. No figures from other plants are published.

This is an anonymised pattern describing how a deployment of this kind is structured. It is not a named customer reference. No customer names, logos, testimonials or measured results are published on this site, because none have been supplied for publication. More on how evidence is handled .

Governance

Privacy and governance in this sector

  • Rules operate on object class and schedule, not identity.
  • Monitoring areas where people work engages employee monitoring obligations and commonly requires consultation before deployment.
  • Analytics supplements guarding, interlocks and permit systems and replaces none of them.
  • Event evidence retention should match the safety and compliance review cycle, with audited access.
  • Recording all events while notifying only exceptions keeps the compliance record complete without alert fatigue.

Frequently asked questions

Will steam and dust cause constant false alarms?

They are the dominant nuisance source in plant environments, which is why rules use object classification and size filtering rather than raw motion, and why known steam and exhaust sources are excluded by region. The residual rate is measured on your own cameras rather than assumed.

Can this be connected to machine safety systems?

No. Camera analytics is a supervisory notification and recording layer. Guarding, interlocks and light curtains are safety-rated functions and must not be replaced or influenced by a video analytic.

How do shift patterns affect the rules?

They largely define them. The same camera typically needs no rule during production, a permit-aware rule during maintenance and a strict rule when the area should be empty. Encoding the shift pattern correctly is most of the configuration work.

Other industries

Author
Gabriel Bamola, Chief Marketing Officer, Ayonix
Technical review
Dr Sadi Vural, Founder and Chief Executive Officer, Ayonix
Published
Last reviewed

Design a factories pilot around your own cameras

Start with a camera readiness review and a log-only baseline, then measure against acceptance criteria we agree before the pilot begins. That is what turns an analytics evaluation into a decision.