License plate recognition SDK

What is Intlab Auto LPR for?

Creating control points with 24/7 optical automatic number plate recognition.

Intlab Auto LPR is a software development kit (SDK) for integrating into third-party applications the ability to detect and recognize plate numbers 24/7 in a broad range of external conditions. The library provides the ability to read license plate numbers in individual images as well as video streams. Using the engine with a video stream provides the best quality recognition since the results received from individual frames captured over the course of a vehicle's motion within the camera's field of view are analyzed and combined to form a single result for each car.

Basic specifications

up to

97%

Recognition accuracy

5-300

Single-frame recognition time

150

km/h

Maximum vehicle speed

All Europe countries

Number of supported countries

All Europe countries

Maximum number of simultaneous recognized countries

Minimum character height

Compatibility

Windows 7, 10

Windows Server

Linux

C/C++

C# (.Net)

Many others

Specifications, system requirements, and the API

Specifications
System requirements
API

Acceptable camera tilt angles (without prior correction of constant tilt angles)	Camera's vertical angle to the plane of the road <= 30°. Camera's horizontal angle to the direction of vehicle motion <= 20°. License plate's roll angle relative to the lower edge of the frame <= 10°.
Vehicle speed	up to 150 km/h
Supported character height	Characters as small as 8 pixels are supported, though a height in the range from 14 to 45 pixels is recommended.
Supported countries	Argentina Belarus Colombia Germany Kazakhstan Poland Russia Spain Ukraine Mode for simultaneous recognition of plate numbers from a 4 countries from the list; Additional mode with a universal recognizer for multi-country recognition of license plates with Latin characters. Regions: all countries in the Europe.
Number of plate numbers simultaneously in the frame	unlimited
Number of video streams	number of video streams depends on the type of license
Supported video signals	individual images or live streams from an analog or digital camera
Licensing system	1 license for each instance of the recognition object, USB dongle Types of licenses: Image (basic license) - recognition on individual frames received from 1 to 10 cameras, engine throughput of up to 10 frames per second, recognition enabled for only one country from the list of supported countries, recognition of one-line and two-line license plate numbers. Various options may also be enabled: simultaneous recognition for several countries, universal recognition; Video (basic license) - recognition on a video stream of up to 30 frames per second from one video image source, recognition enabled for only one country from the list of supported countries, recognition of one-line and two-line license plate numbers. Various options may also be enabled: simultaneous recognition for several countries, universal recognition;
Supported programming languages	SDK can be used in applications in C / C ++, C #, VB.Net, Java and any other programming languages that support calling C functions.
Package contents	SDK distribution package, documentation, examples in C / C++, C#, USB license dongle

Supported operating systems	Windows 7,8,10 (32/64 bit), Windows Server 2008, 2012 (32/64 bit), Linux Ubuntu (64 bit)
Recommended computer configuration:	Core i3 (4th generation desktop CPU or higher) to perform recognition simultaneously on 1-2 video streams. Core i5 (4th generation desktop CPU or higher, 4 cores) to perform recognition simultaneously on 3-4 video streams. Core i7 (4th generation desktop CPU or higher, 8 cores) to perform recognition simultaneously on 5-8 video streams. RAM: 4 Gb or greater.

Engine input

images loaded from a file or passed in a memory buffer in RAW format
real-time video stream passed in a memory buffer in RAW format

Engine settings

Frame resolution, rectangular recognition zone within the frame, minimum and maximum size of a one-line number (the dimensions of two-line numbers are calculated automatically by the engine based on the one-line size), timeout period to produce a recognition result for a car that has stopped or left the field of view, set of allowed license plates (recognition for specific countries, or multi-country mode)

output

Image license
After each frame is processed, the following results are output:

a set of recognition hypotheses, in which each hypothesis contains a UNICODE string representation of the number, a recognition confidence score (hypothesis weight), and the timestamp and image of the frame with the highest confidence score for this hypothesis, and the location of the number within the frame;

Video license
When the engine receives an event indicating that a vehicle has left the frame or stopped, the engine returns the following results:

the timestamp of the first frame where the vehicle was detected;
the timestamp of the last frame where the vehicle was detected;
a set of recognition hypotheses, in which each hypothesis contains a UNICODE string representation of the number, a recognition confidence score (hypothesis weight), and the timestamp and image of the frame with the highest confidence score for this hypothesis, and the location of the number within the frame;
direction of vehicle motion: away from the camera, toward the camera;
array of data for all numbers found as a particular vehicle passes, where each element of the array contains a UNICODE string representation of the number, a recognition confidence score (hypothesis weight), timestamp of the frame with the highest confidence score for this hypothesis, and the location of the number within the frame;

Examples of successfully recognized vehicle number plates

Key advantages

Geography

Broad ability to recognize plate numbers from various countries. The engine was developed from the beginning to recognize plate numbers from most countries equally well, including during simultaneous recognition of license plates from various countries. Our vast and constantly growing test set of more than 8 million images lets us test each version thoroughly to evaluate recognition quality, ensure there are no hidden bugs, and be confident we have produced a high-quality product.

Reliability

Our vast and constantly growing test set of more than 8 million images lets us test each version thoroughly to evaluate recognition quality, ensure there are no hidden bugs, and be confident we have produced a high-quality product.

SIMPLE
TO USE

Для развертывания модуля на объекте требуется минимальное время, так как модуль использует минимальное количество настроек. Установка и настройка модуля требует минимальной квалификации.