Ayonix Video Analytics

Industry

Video analytics for city mobility and public-space management

In short

City video analytics converts existing street and junction cameras into aggregate measurements: vehicle volumes by direction, pedestrian counts, parking occupancy and public-space activity. Outputs are aggregate by design, with plate recognition treated as a separate decision requiring its own lawful basis.

High-angle view of vehicles, cyclists and pedestrians at an urban intersection
Conceptual illustration of an instrumented urban intersection. AI-generated conceptual image, not a customer deployment.

The problem

What smart cities teams are dealing with

  • Traffic volumes are known only from occasional manual surveys that are out of date almost immediately.
  • Installing inductive loops means closing roads and the business case rarely survives the civil works cost.
  • Parking occupancy is unknown in real time, so drivers circulate and congestion worsens.
  • Active-travel investment must be justified but pedestrian and cycle volumes are not measured.
  • Residents and councillors scrutinise camera projects closely and a weak privacy position sinks the scheme.

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 forSmart cities
Location Purpose What matters
Junction approaches Directional vehicle counting Views across the carriageway rather than along it, positioned away from habitual queueing points.
Pedestrian crossings and footways Pedestrian counting Elevated placement; active-travel counts need the same view discipline as vehicle counts.
Surface car parks Area occupancy Height matters more than anything else — a low view cannot see past the front row when the area is full.
Public squares and event spaces Density banding Cover physically meaningful areas and calibrate bands against the city’s own event-safety thresholds.
Controlled access points Plate recognition where lawfully justified A dedicated ANPR camera per lane with controlled approach speed; an overview camera will not deliver an acceptable read rate.

Workflow

From event to acknowledged action

Event workflow

  1. 1 Street camera streams are ingested and monitored for availability.
  2. 2 Vehicles and pedestrians are detected, classified and tracked per view.
  3. 3 Counting lines, occupancy zones and density regions are evaluated per the configured schedule.
  4. 4 Aggregate series are accumulated per interval; no individual record is created.
  5. 5 Series are published to city dashboards and open-data pipelines where appropriate.
  6. 6 Where ANPR is deployed, reads are handled under their own retention and access rules.

Operator workflow

  1. 1 Traffic team reviews directional volumes by interval against the network model.
  2. 2 Parking occupancy is published to entrance signage and city apps in near real time.
  3. 3 Density warnings for event spaces reach the city control room with the live view.
  4. 4 Active-travel counts are exported into the scheme appraisal evidence base.
  5. 5 Any ANPR-driven enforcement follows a separate, documented process with its own authorisation.

Operations

Dashboards, integration and deployment

Dashboard metrics

  • Vehicle volumes by direction and interval
  • Pedestrian and cycle counts by location
  • Parking occupancy by area and time
  • Public-space density bands during events
  • Counting availability by location
  • ANPR read rate by lane where deployed

Integration options

  • Webhook and REST delivery into city traffic-management platforms
  • Open-data export for aggregate mobility series
  • VMS alarm delivery for public-space density warnings
  • ONVIF and RTSP ingestion from existing city camera estates
Compatibility states →

Deployment options

  • Edge processing at the cabinet or pole so only aggregate counts traverse the city network
  • On-premise processing at the city control centre
  • Hybrid deployment with local processing and central aggregation
  • Air-gapped deployment where the traffic control network is isolated
Compare architectures →

Evidence

Deployment pattern

Urban mobility pattern

Directional vehicle and pedestrian counting at junctions, area occupancy in car parks, and density banding in public spaces, delivered as aggregate series to city platforms.

Counting accuracy is validated against manual counts on your own junctions during the pilot. Figures from other cities are not transferable because camera geometry, traffic mix and congestion patterns differ.

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

  • Counting, occupancy and density outputs are aggregate and contain no identity or individual record.
  • Plate recognition is a materially different decision: it processes personal data and requires its own lawful basis, impact assessment, retention rule and access control.
  • Public-space camera analytics attract close scrutiny; publishing the analytics purpose, the data produced and the retention rule is usually the difference between a scheme proceeding and stalling.
  • Aggregate mobility series can be published as open data because they contain no personal data.
  • No behaviour-classification or crowd-prediction analytics is deployed, and claims of either should be treated sceptically.

Frequently asked questions

Can camera counting replace inductive loops?

For operational volume measurement it frequently can, and it avoids road closures by using cameras that often already exist. Where a regulator specifies a calibrated survey instrument, camera counting supports the picture but does not replace the specified equipment.

Does a city analytics scheme require ANPR?

No, and most do not. Volumes, occupancy and density are all aggregate measurements with no plate reading. ANPR is a separate decision with substantially higher privacy obligations and should only be deployed where a specific purpose genuinely requires it.

What can we publish as open data?

Aggregate counts, occupancy percentages and density bands contain no personal data and are commonly published. Plate reads are personal data and must not be published in any form.

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 smart cities 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.