License Plate Recognition (LPR/ANPR): Technology and Use Cases

LPR technology combines specialised imaging hardware with optical character recognition (OCR) software to capture, read, and interpret vehicle number plates. The process happens in several stages: the camera first detects a vehicle in its field of view, then isolates the rectangular plate region from the broader image, applies image processing to enhance contrast and correct skew, segments the individual characters, and finally runs OCR to convert the image into machine-readable text. Modern systems powered by deep learning can complete this entire pipeline in under 100 milliseconds, accurately reading plates on vehicles travelling at highway speeds.

The accuracy of an LPR system depends on several interrelated factors: camera resolution, lens selection, shutter speed, illumination, mounting angle, and the OCR engine itself. A poorly specified system will produce frustrating results — missed reads, misread characters, and excessive false matches. Understanding these variables is essential for any reseller deploying LPR solutions, because the technology is unforgiving of shortcuts in camera selection and installation.

The most critical specification for LPR is not the camera's total megapixel count but the pixel density across the plate. As a rule of thumb, you need a minimum of 3 pixels per character height for reliable OCR, but best practice is to aim for at least 15-20 pixels per character. For a standard Australian number plate where characters are approximately 70mm tall, this means the plate region in the captured image needs to be at least 100-130 pixels tall. A 2MP camera with a tight field of view may outperform an 8MP camera with a wide-angle lens, because what matters is how many pixels are dedicated to the plate, not the total image size.

Australian number plates are manufactured with retroreflective material that bounces infrared (IR) light back to its source. Dedicated LPR cameras exploit this by using powerful 850nm or 940nm IR LEDs synchronised with the camera sensor. The 850nm LEDs produce a faint red glow visible to the human eye, while 940nm LEDs are completely covert. When IR illuminates the plate, the retroreflective characters appear as high-contrast dark text on a brightly glowing background, regardless of ambient lighting conditions. This makes night-time capture as reliable as daytime, provided the IR illuminator has sufficient range for the detection distance.

Shutter speed is equally critical when capturing moving vehicles. A standard surveillance camera might use a shutter speed of 1/30th of a second, which is far too slow for LPR — at 60 km/h, a vehicle moves approximately 55 centimetres during a 1/30s exposure, resulting in severe motion blur. Dedicated LPR cameras use shutter speeds of 1/500s to 1/2000s, freezing the plate image even at highway speeds. The trade-off is that faster shutters let in less light, which is why dedicated IR illumination is essential — the IR LEDs compensate for the reduced exposure time. Many LPR cameras also use a global shutter rather than a rolling shutter to prevent skew distortion on fast-moving plates.

Camera mounting angle has a significant impact on read accuracy. The ideal horizontal angle is no more than 30 degrees off the vehicle's line of travel, and the vertical angle should be kept below 25 degrees. Steeper angles distort the plate characters and reduce the effective pixel density. For car parks and boom gate applications, the camera is typically mounted 1.0-1.5 metres high on a bollard or wall at a distance of 3-5 metres from the capture point. For road-side deployments monitoring faster traffic, cameras are often mounted on poles at 3-5 metres height, looking down the lane at a distance of 15-25 metres. Each camera should monitor a single lane — attempting to cover multiple lanes with one camera inevitably produces poor angles on the outer lanes.

The most common LPR deployment in Australia is parking management. Shopping centres, airports, hospitals, and universities use LPR to automate entry and exit, calculate parking duration, and enforce time limits without physical tickets or boom gates. The system captures the plate on entry, starts a timer, and matches the plate on exit to calculate the fee or verify that the vehicle has not exceeded the free parking window. Integration with payment kiosks or mobile apps allows drivers to pay by entering their plate number, creating a seamless ticketless experience.

Access control is another major use case. In corporate campuses, warehouses, and gated communities, LPR replaces or supplements traditional access methods like swipe cards or remotes. Authorised plates are stored in a whitelist — when a matching plate is detected, the barrier opens automatically. This is especially convenient for logistics operations where delivery vehicles change frequently. Integration with visitor management systems allows pre-registration of guest plates, granting temporary access without issuing physical credentials. Dahua and Hikvision both offer integrated LPR cameras with barrier control outputs, while third-party platforms like Genetec AutoVu provide more sophisticated fleet management and multi-site capabilities.

LPR usage in Australia differs significantly between law enforcement and private entities. Police forces across all states and territories operate mobile and fixed ANPR systems that check plates against databases of stolen vehicles, unregistered vehicles, and persons of interest in real time. These systems operate under specific legislative authority and are exempt from many privacy constraints that apply to private operators. Private use of LPR — by shopping centres, car parks, or body corporates — falls under the Privacy Act 1988 and relevant state surveillance legislation. Private operators must have a clear, lawful purpose for collecting plate data, must display signage informing visitors that LPR is in use, and must have a privacy policy that explains how the data is stored, used, and eventually destroyed.

Successful LPR installation follows a methodical process. Start by defining the capture zone — the exact point where vehicles will be when their plates are read. For boom gate applications, this is straightforward: the vehicle stops at a known position. For free-flow applications like car park driveways or road monitoring, you need to account for vehicle speed and the camera's frame rate to ensure at least 3-5 frames capture the plate. Mount the camera on a stable structure — poles that sway in wind will cause the capture zone to shift. Use a varifocal lens so you can fine-tune the field of view on site. After installation, test with a variety of vehicle types (sedans, SUVs, trucks) and plate types (standard, custom, interstate, trade plates) to verify accuracy across all scenarios the system will encounter.