traffic analytics

Vehicle counting by direction, lane and interval

In short

Vehicle counting applies directional line crossing to tracked vehicles, producing counts by direction and interval at entrances, lanes and access roads. It uses cameras you already have, avoiding the civil works involved in inductive loops, and reports aggregate throughput without reading plates.

Request a pilot Send a sample video

Conceptual illustration of an instrumented urban intersection. AI-generated conceptual image, not a customer deployment.

Scope

What this analytic does, and what it does not

What it detects

Vehicles crossing a counting line in each direction
Throughput per interval per lane or entrance
Peak periods across a day or week
Directional imbalance between inbound and outbound movement

What it does not guarantee

Exact counts in stop-start congestion where vehicles queue across the line
Reliable separation of closely spaced vehicles in a single lane at low frame rates
Vehicle identity or classification beyond broad class
Equivalence with a calibrated traffic survey instrument

The problem

What customers are actually dealing with

These are the situations that lead teams to look at this analytic in the first place.

Access-road volumes are unknown because installing loops means closing the road.
Barrier counts exclude everything that does not pass the barrier, including exits through a separate gate.
Traffic surveys are commissioned for a week every few years and are out of date almost immediately.
Site traffic assumptions in a planning submission cannot be evidenced.

How it works

Detection workflow

Every event carries the rule that produced it, so an operator can see why they were alerted.

1

Ingest

Lane or access-road cameras are read and monitored for availability.
2

Detect and track

Vehicles are detected and tracked so each produces one continuous path.
3

Evaluate crossing

Each track is tested against the counting line and assigned a direction.
4

Aggregate

Crossings accumulate into configurable reporting intervals per lane and direction.
5

Publish

Interval counts are written to the dashboard and made available for export.

Configuration options

Counting line position and geometry per lane or carriageway
Direction labelling consistent across all counted locations
Reporting interval, commonly fifteen or sixty minutes
Region masks to exclude adjacent carriageways or parking not in scope
Minimum track length before a crossing is valid
Optional broad class split where the view supports it

Alert quality

How irrelevant and duplicate alerts are reduced

No video analytic eliminates false alerts. These are the mechanisms that reduce them, and the residual rate is measured on your own cameras.

Track-based counting prevents a single vehicle being counted repeatedly as it crosses the line region.
A minimum track length before the line excludes objects that first appear on the line itself.
Region masking prevents traffic on an adjacent road from contributing to the count.
Reporting availability alongside counts prevents stream gaps being read as genuinely quiet periods.
Congestion across the line remains the dominant error source and is quantified during the pilot.

Prerequisites

Camera requirements and environmental limits

Camera suitability decides more of the outcome than any software setting. These are assessed per camera before commitment.

Camera requirements

A view across the carriageway rather than along it, so vehicles separate at the line
A frame rate of at least 15 fps, higher where vehicle speeds are significant
Sufficient shutter speed to avoid motion blur across the line at the prevailing speed
Adequate night illumination or IR, with exposure that tolerates headlights
Stable mounting, since a shifted line silently changes what is counted

Environmental limitations

Queued traffic straddling the line is the hardest condition and produces the largest errors.
Motorcycles and cycles may be under-counted depending on view and thresholds.
Very high speeds at low frame rates can produce missed crossings.
Counts are aggregate throughput measurements, not a substitute for a calibrated survey instrument where a regulator requires one.

Delivery

Alerts, evidence and where they land

Event and evidence fields

Lane or line name and direction
Interval start and end with timezone
Count in the interval
Broad class split where configured
Camera availability percentage
Site identifier

VMS integration

Vehicle counts are a data series consumed on operational dashboards rather than VMS alarms. Where the same line also requires an out-of-hours crossing alert, a line crossing rule delivers that event separately.

See compatibility states →

Deployment options

Edge processing at the access point so only counts traverse the network
On-premise processing where road cameras already aggregate locally
Multi-site deployment for comparable measurement across estates or campuses

Compare architectures →

Measurement

Dashboard metrics and pilot acceptance criteria

Acceptance thresholds are agreed with you before the pilot starts. This site publishes no benchmark figures, because they do not transfer between sites.

Dashboard metrics

Vehicle counts per direction per interval
Daily and weekly throughput trend
Peak hour per direction
Inbound versus outbound balance over the day
Broad class split where configured
Counting availability per lane

Measurable pilot criteria

Count accuracy against a manual count at free-flow and congested periods
Direction accuracy measured separately from total accuracy
Night performance measured explicitly, including headlight conditions
Behaviour under a scripted queue across the counting line
Interval completeness across a continuous two-week window

Governance

Privacy and governance

Counting produces aggregate totals with no plate reading and no vehicle identity.
Where a camera view includes public road beyond the site, masking limits processing to the area of interest.
Aggregate count series can be retained without retaining imagery.

Frequently asked questions

Can this replace inductive loops?

For operational throughput measurement it frequently can, and it avoids civil works by using existing cameras. Where a regulator requires a calibrated survey instrument, camera counting supports the picture but does not replace the specified equipment.

What happens when traffic queues over the line?

That is the hardest condition and the main source of error, because stationary vehicles straddle the line and tracks merge. Line position is chosen during design to sit away from habitual queueing points, and performance under queueing is measured explicitly during the pilot.

Does it count motorcycles and cycles?

It can, but they are smaller and faster and need their own threshold testing. If they matter to your figures, they are included as a specific acceptance case rather than assumed to be covered.

Keep reading

A controlled pilot establishes what this analytic actually does in your environment, against acceptance criteria we agree before it starts. Send a sample video first if you would rather see an assessment before committing to a pilot.