Works on Blurry Images
Works When Plate is at an Angle
Works When Vehicle is Driving Fast
Works on Plates with Icons
Works in Dark Environment
Works with Multiple Vehicles
Works on Low-Res Images
Works with Motorcycles, Buses, etc.
Works on Plates with 2 Rows
# Get an image
curl -o /tmp/car.jpg https://platerecognizer.com/static/demo.jpg
# Call the API
curl -F 'upload=@/tmp/car.jpg' -H 'Authorization: Token 123456' https://platerecognizer.com/v1/plate-reader
"xmin": 12, "ymin": 84, "ymax": 168, "xmax": 380},
"score": 0.90 }]
See examples of how to interface with our API on this Github . You can call the API on all the files of a directory and analyze the frames of a video.
Sign up to access our API or download our SDK.
Environment Matters. It’s easy to decode a license plate from a high-res photo of a vehicle on a sunny day. But we know that in reality, the weather and other conditions are not always ideal. We have been relentlessly enhancing our algorithms to support various “real-life” factors, such as sun glare, blurry images, fast vehicles, night-time, and many more.
Self-Learning. Our ANPR engine is like a teenager, it continues to learn. This means we can fine-tune our engine to help meet your specific needs, such as a certain camera angle or lighting condition. See how we worked to improve our !
Location-Optimized. Let’s face it–not all plates are the same. Some have two rows of text. Some have icons at the beginning, middle or end of the plate. Some contain fancy fonts. Thankfully, our LPR engine supports them all and is tuned specifically for your region.
Five is Better than One. Unlike other ANPR providers, we can return up to 5 decoded license plates from one image. We achieve this by utilizing two distinct neural networks. One identifies all the photos of license plates from an image and the other decodes each character of the plate.
Automatically open the parking gate based on license plate.
Collect toll automatically based on plate number with the need for RFID!