High-Speed ANPR and Vehicle Occupancy Detection

Tolling and access-control infrastructure is moving away from stop-and-pay booths toward free-flow systems, though the underlying technology mix varies by region. Germany, Belgium, and Lithuania toll roads use satellite-based GNSS tracking for trucks with no RFID tag required; Italy also runs a microwave tag system; France and Spain use 5G-based RFID. No single technology has become a global standard, and every one of these systems still relies on cameras for collecting revenue from non-registered or non-compliant vehicles by plate.

Japan's ETC network and Italy's Autostrade system use gated ANPR pre-validation so a barrier can open before the vehicle arrives. London's congestion-pricing zone is a further example: it relies on NIR-illuminated ANPR cameras to monitor and price traffic through central London, and the same imaging system identifies electric vehicles for a 25% Auto Pay discount on the daily charge.

The optical performance of these cameras is not a peripheral engineering detail; it is tied directly to outcomes that researchers are actively measuring. A 2026 study in MDPI’s (Multidisciplinary Digital Publishing Institute) Infrastructures examining a national motorway's transition from barrier-based tolling to free-flow tolling found that removing physical toll plazas cut vehicle queuing by roughly 79.5%, was projected to reduce accidents in toll areas by approximately 50%, and reduced CO2-equivalent emissions at toll locations by up to 80% for light vehicles and 85% for heavy vehicles, alongside reductions in PM10 particulate emissions, primarily by eliminating the stop-and-accelerate cycle at plazas (1). These effects depend entirely on the camera correctly identifying the vehicle on the first and only pass; a missed or misread plate at highway speed cannot be recovered the way it can at a gated toll booth. This is why the optical system, and not just the recognition software, determines whether a free-flow deployment delivers on its safety and environmental potential.

Designing a reliable ANPR system begins with defining imaging requirements rather than selecting hardware. Each camera installation location presents unique challenges based on roadway geometry, vehicle speed, mounting location, lighting conditions, and regulatory requirements. Successful high-speed ANPR systems, however, must meet several fundamental optical requirements.

1. How Much Plate Resolution Is Enough?

The optical system must provide enough pixel density across the license plate for reliable Optical Character Recognition (OCR). Most ANPR systems target approximately 100-150 pixels across the plate, with higher values providing margin for motion, plate angle variation, image compression, and environmental plate degradation. Achieving this depends on the combined selection of camera resolution, sensor format, lens focal length, working distance, and field of view.

High-speed ANPR requires balancing competing optical requirements: sufficient plate resolution, low motion blur, accurate focus, and reliable operation under variable lighting. The table below summarizes where common approaches fall short in real deployments.

ANPR system commissioning

Large tolling systems must have every camera field of view and focus adjusted for optimal performance. This can be a time-consuming process. Installing cameras where each lens has a pre-defined curve relating best focus to field of view can greatly improve the time required to optimize the image. For motorized lenses, any change in field of view will require a focus motor adjustment so having a camera that will automatically correct the focus will save time and reduce frustration for the technician.

Designing an optical system for high-speed ANPR is fundamentally an image-formation problem. Unlike general surveillance imaging, ANPR performance is judged by a binary outcome (whether the plate is correctly recognized), so optical performance at the license plate location is the primary design constraint.

1. What Determines the Required Working Distance?

Long-range tolling and enforcement systems often require telephoto optics because the camera cannot be positioned close to the vehicle path. Key factors include camera mounting height and offset from the lane, the range of vehicle heights and plate positions expected on that lane, lane width and number of lanes, the capture zone, and the required recognition distance.

2. Why Does Long Working Distance Reduce Motion Blur?

Motion blur depends on the motion of the image of the plate across the sensor during exposure, not on vehicle speed alone. For a camera viewing the vehicle at an oblique or cross-lane angle, image motion is approximately the perpendicular component of vehicle speed divided by working distance. This is an approximation, and the perpendicular-velocity assumption matters: a gantry-mounted camera looking down the roadway sees vehicles approaching nearly head-on, where the perpendicular component (and therefore blur) is small regardless of distance; a roadside camera viewing across the lanes sees a larger perpendicular component that shrinks as working distance increases.

In practice, the more consistent benefit of a long working distance is time, not just angle. At 100 km/h (27.8 m/s), a vehicle traverses a 50 m band in about 1.8 seconds. At low perpendicular travel angle such as a long telephoto lens, this can all be in the camera’s acceptable image capture range. At a closer distance, the camera will be at a steeper angle to the travel direction and the acceptable capture range could be a fraction of a second. That large difference in available time is what actually reduces motion-blur risk in practice: it gives the camera and NIR illumination system a wide window to trigger exposure, hold focus, and select the sharpest frame, rather than depending on a single, precisely timed instant.

Distance creates its own tradeoff, however: the license plate occupies fewer pixels as working distance increases. Long-range ANPR therefore requires telephoto optics that provide longer focal length for plate magnification, short exposure times for motion control, and focused NIR capability for consistent nighttime imaging.

3. How Many Pixels Across a License Plate Are Required?

ANPR recognition accuracy depends strongly on the number of pixels covering the plate and individual characters. Basic detection typically needs roughly 50-80 pixels across the plate width, reliable OCR recognition needs approximately 100-150 pixels, and high-confidence highway ANPR should target 150-200 or more pixels. Designing for approximately 150 pixels or more provides margin for vehicle motion, plate angle variation, lighting changes, compression artifacts, dirty or damaged plates, and OCR processing variation. Pixel count on the plate scales with plate size, focal length, and sensor resolution, and scales inversely with working distance and sensor size.

Plate geometry differs meaningfully by region and should be designed for separately. US plates measure a fixed 300 × 150 mm. EU plates are wider and shorter — 520 × 110 mm is the dominant standard across most member states, though some countries use narrower formats down to roughly 340 mm, so EU width should be treated as a 340–520 mm range with 520 mm as the representative highway case. The EU plate's extra width is also spread across more characters (typically 7–8, versus 6–7 on a US plate) packed into a shorter plate height, so its characters are smaller and more tightly spaced than the total plate-width pixel count alone would suggest. For system design, the whole-plate pixel figures below should be treated as a first-order check; per-character pixel resolution — and extra margin for the denser EU layout — is the more reliable metric where EU-format plates are in scope.

Worked example: at 4K resolution (3,840 horizontal pixels) and a field of view of approximately 12 m (wide enough to span three highway lanes), a system delivers roughly 320 pixels per meter of scene width. Applied to a 300 mm US plate, that works out to roughly 96 pixels across the plate — just below the ~100-pixel figure commonly cited as a practical minimum, and adequate only for basic detection at this FOV. Applied to a 520 mm EU plate, the same setup yields roughly 166 pixels across the plate — clearing the reliable-OCR threshold on a whole-plate basis, but with more characters packed into that width and a shorter plate height, per-character resolution is tighter than the raw number implies, so this should be read as a marginal rather than comfortable result.

Zooming the same lens toward its telephoto end to cover a single lane (roughly 4 m FOV) at the same working distance increases density proportionally to about 960 pixels per meter. That yields roughly 288 pixels across a US plate and roughly 499 pixels across an EU plate — comfortably above the 150–200 pixel high-confidence target in both cases, with the EU plate's larger character count now well covered too. This is the practical tradeoff a motorized zoom lens is built to manage: wide-FOV, multi-lane coverage trades away plate (and per-character) resolution, and a longer focal length recovers it on demand, in the field, without swapping optics — and the EU case, being the denser and more demanding format, is the one to validate against when designing for mixed-region deployments.

4. How Should Focal Length Be Selected?

Focal length selection should begin with required field of view and plate pixel density, not simply vehicle distance. A shorter focal length provides a wider field of view, greater scene coverage, and easier vehicle acquisition, but reduces plate magnification. A longer focal length provides higher plate pixel density, greater recognition distance, and improved character resolution, but reduces field of view, vehicle positioning margin, and alignment tolerance. A motorized telephoto lens, such as a 12-50 mm design, allows the optical system to be adjusted after installation: the wide end for roadway coverage and easier vehicle capture, the long end for increased magnification and long-range plate recognition.

5. How Does NIR Imaging Improve ANPR and Occupancy-Detection Performance?

Near-infrared imaging creates a controlled lighting environment that reduces dependence on uncontrolled daylight illumination. Many license plates contain reflective materials that respond strongly to NIR wavelengths, and combining NIR illumination with a NIR-corrected lens provides higher plate-to-character contrast, reduced sensitivity to sunlight changes, improved nighttime operation, and more consistent image quality in changing weather. The same NIR path also improves occupancy counting through tinted glass. A lens designed only for visible wavelengths may experience focus shift when switching to NIR, because different wavelengths focus at different optical positions. ANPR optics should maintain correction across approximately 430-940 nm to support daylight imaging, white LED illumination, 850 nm NIR illumination, and 940 nm NIR illumination; this correction becomes increasingly important in telephoto systems, where chromatic focus errors are more noticeable.

6. Why is Back Focal Length Calibration Required?

Back Focal Length (BFL) calibration is required because the relationship between a lens and camera sensor varies between individual systems. A motorized zoom lens contains a predefined zoom/focus tracking model describing how focus changes with focal length, but this model represents average production behavior. Individual assemblies vary due to mount tolerances, mechanical stack-up variation, and sensor position. These differences are most significant at the telephoto end, where focus tolerance is smallest. BFL calibration aligns the optical model with the actual camera-lens assembly, which matters for ANPR because recognition distance depends on maximum image sharpness, OCR relies on fine character edges, telephoto systems have limited depth of field, and motorized zoom requires repeatable focus control. BFL calibration is done at the manufacturer so the camera installer can quickly and accurately find the optimal focus in the exact setup being used.

A high-speed ANPR system should be designed around measurable optical targets. The goal is not simply capturing a vehicle image, but delivering a sharp, high-contrast plate (and, where required, cabin) image that provides reliable information for OCR and occupancy algorithms under real-world conditions.

1. License plate pixel density: at least 150 pixels across the plate for high-confidence ANPR, with sufficient character height for OCR under motion, angle, and lighting variation.

2. Optical resolution: support 4K and multi-megapixel ANPR cameras while maintaining high contrast resolution across the full image area and at long focal lengths. Lens resolving power (spot size) must match or exceed the pixel size of the sensor, so the lens itself is never the limiting factor and the resulting image stays crisp rather than fuzzy.

3. Long-range imaging capability: support long working distances while maintaining plate resolution, with sufficient magnification and without excessive field-of-view limitations.

4.Motion blur control: minimize plate blur and preserve character edges at highway speeds through short exposure times; accomplished with a fast lens that allows high light throughput, controlled illumination, and high-speed camera triggering.

5. Visible-to-NIR optical correction: maintain focus and resolution from approximately 430-940 nm, supporting NIR illumination without focus shift across daylight, low light, active infrared illumination, and weather conditions.

6. Zoom tracking accuracy: provide repeatable positioning and compensate for lens-camera variation via BFL calibration to allow quick and easy camera setup and optimally focused image.

7. Environmental reliability: maintain optical performance during 24/7 outdoor operation, accounting for temperature-induced focus changes, mechanical stability, and vibration.

8. Occupancy detection support: maintain sufficient NIR contrast and resolution to count occupants through tinted or reflective windshields, using the same optical path as nighttime ANPR.

9. System-level optimization: maximize recognition accuracy while minimizing processing requirements, image correction overhead, bandwidth demands, and installation complexity.

A practical high-speed ANPR implementation requires optics that combine long-range magnification, remote configurability, high resolution, and reliable focus control. Theia Technologies' TL1250P-IQ, a 12-50 mm motorized telephoto lens, is one example built to this specification; it provides a flexible optical platform for applications ranging from long straight-road tolling lanes to shorter-range access control and carpool-lane validation. By combining motorized zoom, focus, iris, and filter control, one lens can support multiple roadway geometries without physical replacement or manual refocusing after installation. Its F/1.8 aperture supports shorter exposure times in low-light conditions, helping maintain image sharpness at long working distances.

In deployed systems, a single TL1250P-IQ lens has been used to span four lanes of traffic while monitoring more than 70 m of roadway, and to deliver long-range LPR/ANPR at distances up to 140 m as well as 500 ft in mobile enforcement cameras (vendor-reported deployment figures. The TL1250P-IQ pairs with the MCR IQ motor control board so focus and field of view can be set and adjusted entirely after the gantry is installed, without stopping traffic to access the camera.

Lens Specifications

• 12-50 mm motorized telephoto zoom range
• Motorized zoom, focus, P-iris, and filter control
• F/1.8 @ 12mm - F/2.4 @ 50mm to close
• 4K / 12 MP compatible; resolution performance designed for a 1.55u size pixel delivering up to 300 lp/mm
• Visible-to-NIR correction from approximately 430-940 nm, with internal switchable filters including IR cut/visible only, 850nm notch and 940nm long pass.
• Adjustable field of view: approximately 9 degrees (long-range highway) to 36 degrees (shorter-range tolling or access)
• Supports sensors up to 1/1.7" format
• CS mount or D25 board mount
• Compact design: <65mm total track length (TTL), 50 mm lens length, approximately 74 g mass
• Integrates with Theia's MCR IQ motor control board

Software and Control Stack

The IQ Lens System combines motorized optical control with lens-specific calibration data to enable repeatable optical configuration. It includes a JSON calibration file (average zoom/focus tracking curve, distortion characteristics, aperture information, relative illumination data), an SDK providing GUI-based control and Python or C++ integration, and the MCR IQ motor control board supporting zoom, focus, iris, and filter switching over USB, I2C, and UART communication; RS485 is also in development.

Videos

High-performance ANPR optics improve recognition reliability, but final system performance depends on the interaction between the lens, camera, illumination, processing pipeline, and installation environment.

1. How Do Field of View and Recognition Distance Trade Off?

Wider field of view improves multi-lane coverage, alignment tolerance, and vehicle acquisition, while longer focal length improves plate magnification and recognition distance at the cost of scene coverage. The optimal balance depends on roadway geometry, camera mounting height, lane width, and required capture zone. At telephoto ranges, small optical variations - such as focus error, lens alignment, sensor positioning, and chromatic aberration - become increasingly important, so high-resolution ANPR systems require careful matching between lens, sensor, and camera platform.

2. How Sensitive Are Depth of Field and Focus?