What Is ANPR? Automatic Number Plate Recognition

ANPR (Automatic Number Plate Recognition) is camera-based software that reads a vehicle’s license plate from live video and converts it into searchable text in real time. It is the same underlying technology most people in the United States call LPR (License Plate Recognition): a camera captures the plate, optical character recognition (OCR) isolates and reads the characters, and the resulting text is matched against a watchlist, a gate log, or a time-stamped record of every vehicle that entered a site.

What ANPR Actually Does

ANPR does not identify a driver or a vehicle’s owner on its own. It reads the plate character string and hands that string to whatever system is watching for it, whether that is a parking gate, a police alert list, or a yard-management log at a port. The camera hardware itself typically follows the same interoperability standards, such as those published by ONVIF, that govern IP camera compatibility more broadly, which is why ANPR is usually added as a capability on top of existing camera infrastructure rather than requiring an entirely separate system.

The Three-Step Read Pipeline

Every ANPR system, regardless of vendor, runs the same three steps in sequence. First, capture: a camera fitted with either a fixed lens tuned to a specific lane width or a varifocal lens photographs the plate, usually triggered by a loop detector, radar, or continuous frame analysis rather than a human pressing a button. Second, character isolation: OCR software locates the plate region within the frame, corrects for skew, and segments the individual characters, a step that runs in well under a second on modern edge hardware. Third, matching: the resulting text string is compared against whatever list or log the deployment is built for, and only the plate text, not the underlying photo, typically travels onward into the log unless the site is also configured to retain images.

Where ANPR Is Used

Common Deployment Points

• Port and terminal gates, matching incoming trucks against a manifest
• Parking facilities, for touchless entry and exit
• Perimeter checkpoints at gated commercial or industrial sites
• Toll roads and congestion zones, reading plates at highway speed without requiring a vehicle to slow down

The capture conditions differ sharply between these use cases. A gate deployment typically reads a plate at under 15 mph with the vehicle nearly stationary, while a toll or free-flow congestion-zone deployment has to read the same plate accurately at speeds up to roughly 70–100 mph, which is why the two use cases often specify different camera and lens configurations even though the underlying OCR software is the same.

Camera Specifications That Determine Read Accuracy

Resolution, Frame Rate, and Low-Light Performance

Read accuracy is set largely by whether the camera captures enough pixels across the plate itself, not by the OCR software alone. A plate that fills too small a portion of the frame, because the camera is mounted too far away or the resolution is too low, will not resolve into individual characters no matter how good the recognition algorithm is. Most dedicated ANPR cameras run at 1080p to 4K (2 MP to 8 MP) and rely on either a dedicated infrared illuminator or a global-shutter sensor to keep a fast-moving plate sharp after dark, since a standard rolling-shutter sensor can smear a plate that is moving at speed under low light.

Environmental and Mounting Standards

Because ANPR cameras are almost always mounted outdoors, on a gate arm, a road gantry, or a pole, they are commonly rated IP66 or IP67 under the IEC 60529 ingress-protection standard, and powered over Ethernet under IEEE 802.3af (15.4 W) or, for units with a heated housing in cold climates, 802.3at (30 W). A standard Cat6 PoE run is limited to 100 meters before signal loss requires a mid-span extender or a media converter, a limit that matters at ports and large lots where the gate is often well past that distance from the nearest switch.

What Limits Accuracy in Practice

Read accuracy drops with a dirty or obscured plate, extreme camera angles, poor lighting at night, and non-standard plate formats crossing between regions or countries. A site depending on ANPR for a security decision, rather than a convenience feature, should test it against its own worst-case lighting and angle conditions before relying on it, since a plate captured at more than roughly 30 degrees off-axis from the camera is materially harder to resolve than one captured close to head-on.

What Is an ANPR System?

An ANPR system is the full chain that turns a vehicle on camera into a usable record: one or more cameras, OCR software that reads the plate, a list or log to match it against, and the network and power that connect them. The camera alone is not the system. What the plate text is matched against, and what happens next, decides whether a site gets a barrier that opens, a gate log, or an alert.

Parts of an ANPR System

• Camera and lens, fixed to a lane width or varifocal, usually with an infrared illuminator for night reads
• OCR software that finds the plate in the frame, corrects skew and reads the characters
• A matching list: a watchlist, a permit list, a manifest or a time-stamped entry log
• Network and power, commonly PoE under IEEE 802.3af or 802.3at, with a 100 m limit per Cat6 run
• Retention rules for stored reads, set by the operator’s own policy, commonly 30 to 90 days

How an ANPR System Works

The system runs three steps for every vehicle: capture the plate, read the characters, and match the text. The detailed version, including what triggers the camera, is in the three-step read pipeline above.

ANPR System Price: What Drives the Cost

There is no single price for an ANPR system, because the cost follows the site rather than the product. The same software can run one gate lane or a hundred cameras, and the number of lanes, the camera type and the read conditions move the total more than any list price does.

Factors That Change the Cost

• Number of lanes and camera positions that need a plate read
• Dedicated ANPR cameras versus software running on cameras the site already owns
• Capture speed: a gate reading vehicles under 15 mph is a different build from a free-flow lane at highway speed
• Cable distance, since a PoE run beyond 100 m needs a mid-span extender or media converter
• Storage and the retention window for plate reads and any retained images
• Integration with barriers, a yard-management log, a parking system or an alert list

Dedicated Cameras or Software on Existing Cameras

The largest single choice is whether to buy a plate-reading camera for each lane or to add ANPR as software on cameras already installed. Argu runs on existing camera feeds at 1080p to 4K and 15–30 fps over RTSP or ONVIF Profile S, so a site with suitable cameras may not need a new camera at each lane. Whether a given camera is suitable depends on its position, lighting and angle, which is why a test on the site’s own footage comes before any quote.

ANPR System for Tolls and Highways

An ANPR system for tolls reads plates in free-flow traffic, without asking vehicles to slow down. Toll and congestion-zone reads happen at speeds up to roughly 70–100 mph, so these sites typically use cameras purpose-built for that capture distance and vehicle speed.

Gate Reads Compared With Free-Flow Reads

A gate or car park reads a nearly stationary vehicle at under 15 mph, which is why software on an existing camera can work there. A highway lane gives the camera a fraction of a second to freeze the plate, and that is where lens choice, shutter speed and mounting height decide the result.

ANPR System in the UK

In the UK and Ireland the technology is called ANPR, automatic number plate recognition. The United States calls the same thing LPR, license plate recognition. The camera and OCR pipeline are the same, and the term only changes the search results you see.

ANPR in Car Parks

Car parks are the most common UK use of ANPR. Two reads, entry and exit, give the time on site without a ticket. Plate data counts as personal data under UK GDPR, so retention limits apply from the first read. The full walkthrough is in the guide to ANPR parking.

How to Choose an ANPR System or ANPR Camera

There is no single best ANPR system or best ANPR camera, because the right answer depends on the lane, the speed, the lighting and what the plate read has to trigger. Judge each option against your own site conditions instead of a ranking.

A Short Checklist

• Test on your own footage at night, at the real mounting angle, instead of relying on a general accuracy claim
• Confirm the camera supports ONVIF Profile S or Profile T if it must work with existing recording software
• Check the outdoor rating (IP66 or IP67) and the power budget for the cable run
• Ask where plate data is stored and for how long, and whether on-premise deployment is available
• Check what the system reads besides the plate. Argu also captures vehicle color, make and a visible fleet logo in the same frame, which gives a second way to confirm a vehicle when the plate is dirty or partly hidden

For a comparison of the two camera approaches, see license plate recognition cameras.

ANPR and Data Protection

A license plate is not inherently anonymous. Under the EU’s GDPR Article 4, personal data includes information that can identify a natural person indirectly, and a plate number that a records system can trace back to a registered owner falls within that definition, which is why many ANPR deployments in the EU are treated as processing personal data from the point of capture, not only once a name is attached. In practice, most private operators purge non-alert plate reads on a set retention window, commonly in the range of 30 to 90 days, a figure set by the operator’s own policy rather than by a single fixed rule, since no one national law sets a universal retention period for every use case. Sites that also handle camera or component sourcing for a public-facing or government-adjacent deployment should separately confirm compliance with rules like NDAA Section 889, which restricts certain named vendors from federally funded systems, a check that is independent of ANPR itself but often runs alongside it. Car park operators can see how ANPR parking uses plate reads for entry, exit, and permits.

Frequently Asked Questions

What is the difference between ANPR and LPR?

ANPR and LPR describe the same technology. ANPR (Automatic Number Plate Recognition) is the term used more often in the UK, Europe, and much of the rest of the world, while LPR (License Plate Recognition) is the more common term in the United States. Both refer to a camera and OCR pipeline that reads a vehicle plate and converts it to text.

How accurate is ANPR in real-world conditions?

Accuracy depends heavily on the specific deployment: camera resolution, mounting angle, lighting, and plate condition all move the read rate up or down independently of the OCR software itself. A site evaluating ANPR for a security-critical decision should test the system against its own worst-case conditions, such as night capture at the actual mounting angle, rather than relying on a vendor’s general accuracy claim.

Does ANPR require special cameras, or can it run on existing cameras?

ANPR can often run as a software capability added to existing IP cameras that already meet baseline resolution and frame-rate requirements, provided they support a standard interoperability profile such as ONVIF Profile S or Profile T. High-speed or long-range deployments, such as free-flow tolling, typically still need a camera purpose-built for that capture distance and vehicle speed.

Is ANPR data considered personal data under privacy law?

Under GDPR Article 4, a plate read that can be linked to a registered vehicle owner is treated as personal data, which brings retention limits, access rights, and processing obligations into play. Comparable state-level privacy laws in the United States, such as the CCPA, apply similar identifiability-based reasoning to automated license plate reader data.

Can ANPR read plates at highway speed?

Yes, with the right camera and lens configuration. Free-flow tolling and congestion-zone systems are built specifically to read plates at speeds up to roughly 70–100 mph without requiring a vehicle to slow down, which is a materially different engineering problem than a gate deployment reading a nearly stationary vehicle.

What is an ANPR system?

An ANPR system is the set of cameras, OCR software, matching list and network that reads a vehicle’s number plate from live video and turns it into text. The text is matched against a watchlist, permit list or entry log to trigger an alert, a barrier or a record.

How does an ANPR system work?

It captures the plate with a camera, isolates and reads the characters with OCR software, then matches the text against a list. Only the plate text usually travels onward unless the site is set to keep images.

How much does an ANPR system cost?

The cost depends on the number of lanes, whether dedicated ANPR cameras are needed or existing cameras can be reused, the capture speed, cable distances, storage and integrations. A fixed price does not exist, so a test on the site’s own footage comes before a quote.

What is the best ANPR camera?

No single camera is best for every site. A narrow-view camera with a fast shutter and infrared illuminator suits a high-traffic gate or toll lane, while software on an existing camera suits secondary areas. Test the shortlist at your own mounting angle and in night conditions.

Can an ANPR system be used for tolls?

Yes. Free-flow tolling reads plates at speeds up to roughly 70–100 mph, which usually calls for cameras built for that distance and speed rather than a general surveillance camera.

Argu’s OCR for ports module and LPR cameras agent apply this same plate-reading approach on top of a site’s existing gate cameras, so a terminal or facility does not need to install separate ANPR hardware to add the capability. It is one of the plate- and vehicle-facing tools built for smart city deployments, alongside dedicated parking lot surveillance for touchless entry and exit.

Last updated: October 2026