Leading the Edge in Deep Learning License Plate Recognition
LILIN AIDA ANPR Software is powered by advanced deep learning artificial intelligence algorithms, replacing traditional optical character recognition (OCR) and computer vision methods. By learning from massive datasets of real-world license plates, AIDA ANPR delivers outstanding recognition accuracy and stability under extreme lighting conditions, bad weather (rain, fog, haze), steep angles, and high-speed motion. The engine is capable of processing 12 to 48 recognition cycles per second, supporting vehicles traveling at speeds up to 200 km/h, providing a robust core for next-generation smart traffic and automation control.
Core Advantages & Features
- Up to 99.6% Accuracy: Exceptional recognition rates even under high-speed traffic and multi-lane setups.
- Patented Background Filtering: Patented algorithm to automatically filter out ambient background texts, billboards, and road signs, focusing solely on the license plate characters.
- Steep Installation Angles: Supports steep vertical (up to 30 degrees) and horizontal (up to 30 degrees) angles.
- Multi-Lane & Global Support: Recognizes plates from various countries (USA, Europe, Middle East, Japan, Southeast Asia, Taiwan, etc.) across multiple lanes simultaneously, requiring only a minimum plate width of 100 pixels.
System Architecture & Application
Open SDK & API Integration Solutions
AIDA ANPR Software is built with system integrators and software developers in mind. The system operates as a background server through GyNet.exe, and the core recognition process is exposed via AIEngine.exe using a standard HTTP interface. Third-party Video Management Software (VMS) or access control platforms can send a simple HTTP POST request with a JPEG image payload to the target Port (default: 127.0.0.1:8592/sendjpeg) and receive JSON-formatted recognition results (including plate characters, confidence score, and bounding box coordinates) within milliseconds.
This architecture allows for seamless integration with LILIN Navigator VMS, parking management platforms, smart virtual fences, and automated gate controllers without complex SDK integration, dramatically reducing project deployment time.
OS & Software Environment
- Operating System: Windows 10 (Version 1903 or above)
- Graphics Components: DirectX 12
- AI Runtimes: NVIDIA GPU Driver (CUDA 10.2 and cuDNN 10.2 or above recommended) or Intel Movidius / OpenVINO Toolkit.
Recommended Hardware
- CPU: Intel Core i7 or above
- System Memory: 16 GB RAM or above
- Acceleration Hardware: NVIDIA Graphics Card (CUDA supported) or Intel HD Graphics 630+, UP AI Core XM 2280, or Intel Movidius NCS2 x 2.