Traffic & mobility / Flagship demonstration

Know what is using the road. And how.

Automated vehicle counts, classification and speed monitoring for roads, junctions and car parks. Build evidence for transport planning, traffic calming and active travel without reading number plates.

Discuss your project Explore the demonstration films

One road. Two directions. Separate counts.

Each crossing records a direction and object class. The arrows show permitted movement, not observed vehicles. Speed measurement uses a separately calibrated zone. Illustrative, not live data.

The measurement brief

Type. Direction. Time. Context.

A count is more useful when you know what crossed the line, which way it travelled and what was happening around it.

Vehicle classification

Count cars, buses, motorcycles and bicycles. Delivery vans, HGVs, trucks and service vehicles use a unified utility class; pedestrian detection supports active travel studies.

Directional flow

Examine peak hours, historical patterns, junction turning movements and queue lengths. Combine current dashboards with repeatable reporting periods.

Calibrated speed zones

Measure speeds in a calibrated zone, set thresholds and review traffic-calming interventions. Optional speed-event snapshots provide evidence without continuous recording.

Measure the question you have.

A road survey needs counts by type, direction and time. A junction study needs turning movements. A speed study needs a suitable camera view and calibration, not just a line drawn on a picture.

Reports and CSV exports support planning, compliance work and further analysis. Standard analytics do not read number plates or recognise faces.

See how counts become events

Field conditions

A model built for the view from above.

The traffic model is custom-trained for UK monitoring conditions, rather than relying only on eye-level photographs.

Elevated and oblique views

Perspective-matched training covers vehicle roof profiles, along-road views, side-on positions and oblique junction angles. Distant vehicles occupy a smaller part of the frame.

A consistent utility class

Vans, trucks and service vehicles are grouped consistently rather than relying on incompatible commercial-vehicle labels across datasets.

Partial visibility and tracking

Training addresses partly occluded vehicles at barriers and frame edges. Stable multi-object tracks help reduce duplicate detections and improve count continuity.

Assess the real operating conditions

Daylight, overcast skies, dawn, dusk and street-lit night, including mixed sodium and LED lighting.
Dry roads, rain, fog, mist, wet-road reflections, glare and backlight.
Elevated installations, including typical lamppost heights of 5-10 metres, in urban and A-road settings.
Free-flowing traffic, rush-hour congestion, queues and mixed vehicles, cyclists and pedestrians.

Lighting, weather, camera angle and occlusion still affect accuracy. A pilot establishes the operating envelope for your site; this is not a claim that a camera sees through obstructions.

See the original work

Edinburgh, in daylight and after dark.

Six original demonstrations filmed from a drone at intersection height, simulating lamppost viewpoints. These are demonstration recordings, not evidence of permanent installations at the locations.

Film 01 / Day