Videoloft’s LPR model

Our LPR feature is designed to detect, recognize and analyze vehicle license plates with precision.

How it works

During the day

Sequence of video frames showing a vehicle driving down a residential street as the system analyzes movement across multiple frames.

Multiple still frames are extracted and analyzed to detect a variety of vehicle types, including cars, buses, trucks, and motorcycles.

Roadway detection zone highlighted in yellow to define the area where license plates are captured.

The detected vehicle results are then checked against user-defined detection zones. 

License plate clearly captured and centered within the frame after successful recognition.
We select the frame with the best chance of delivering an accurate plate read and send it to third party plate reading platform Plate Recognizer.
Videoloft's license plate recognition dashboard within the Video Management System, displaying LPR data such like time, plate registrations, make, model, colour, plate location, and the camera it was captured on.
Results are displayed in the LPR dashboard for each search, investigation and evidence download.

At night

Instead of detecting the entire vehicle – which is difficult in IR mode due to headlight glare – we detect the rectangular shape of the license plate and pass that directly to Plate Recognizer.

Nighttime security camera view of a car approaching a gated parking entrance, with headlights on and the license plate visible for automatic recognition.

How its trained

Thousands of cameras from Videoloft customers have been opted in to Videoloft’s analytics training programme, so the model is trained with real customer CCTV footage from cameras globally –  ensuring accuracy in real world angles, lighting, and environments.

Releases

Each release is tested extensively and benchmarked. Below you’ll find version history, accuracy scores, and known limitations.

V16 Feb 26 97% accuracy

Release Date February 2026

Summary Current released version

Accuracy results

Metric Accuracy
Highway/road cameras
96% (average across single and dual lane)
Stationary vehicle cameras
(e.g. car park entrances)
98%
Overall vehicle detection accuracy
97%
We test each version of our model against a set of real customer videos. Let’s explain what these numbers mean…
 
  • Release V16 detected 96% of vehicles on highway/road cameras.
  • Release V16 detected 98% of vehicles at stationary check points.
  • Therefore the overall vehicle detection accuracy is 97%.

Images of these vehicles are then fed into Plate Recognizer, you can read about their License Plate Detection accuracy here.

AI detections are not always 100% accurate, so it’s important to be aware of their limitations. The example images below are AI-generated to protect customer data.

Known limitations with our vehicle detection

At night, when cameras operate in IR mode, headlight glare can make it difficult to detect vehicles. Instead, the system switches to a different method and detects the shape of the license plate.

Known limitations with Plate Recognizer plate reading

Overexposed vehicle headlights at night causing glare and reduced visibility for license plate recognition.
The make, model, and color information from Plate Recognizer can vary in accuracy, so we present it as BETA and include a confidence score.

hello@videoloft.net

Toll-free US: (855) 737-0237
UK: +44 203 9870133

+44 753 280 1187

Try it for free

For more information, contact us directly or fill in your details using our quick form and a representative of our global sales team will be in touch.

Your details

We value your privacy and take data protection seriously. We will never share or sell your personal data to third parties. Your trust is important to us, and we’re committed to safeguarding your information.