Vision

Video

Smart city

Transportation

Partner

Number-Plate Recognition (ANPR) based on Vision AI by Irida Labs

Vision

Video

Partner

Find out more

As cities and infrastructure evolve, managing vehicle traffic efficiently is a priority for municipalities, businesses, and residential applications. Traditional ANPR solutions rely on cloud-based, which often face privacy, high cost, latency, and adaptability issues. Customers require a real-time, cost-effective, and scalable solution for applications like parking management, toll booths, and traffic monitoring. Additionally, smart building and smart home applications, such as automated smart parking, vehicle control, smart garage doors, and gated community access control, can also benefit from advanced ANPR solutions.

Irida Labs' Edge AI-powered ANPR solution overcomes these challenges by processing data directly on the edge device, utilizing STMicroelectronics STM32 MCUs. It provides accurate recognition, as of today, of European car registration plates (but can be extended on demand for any country) while operating effectively under multiple weather scenarios and 300-20,000 Lux lighting conditions.

Image processing is performed with a frame rate of 3.1-6.0 FPS (end-to-end). Seamless integration via API/MQTT/HTTP/SERIAL enables applications such as vehicle-based access control, private parking management, and smart tolling, while reducing operational costs by approximately 40% compared to existing solutions.

By eliminating the need for expensive cloud computation and complex server setups, this solution enables municipalities, businesses, and residential communities to enhance vehicle monitoring with minimal investment. Its optimized performance for controlled conditions, combined with real-time processing, makes it a valuable tool for smart city, smart building, and smart home applications.

We'll dig more into the implementation details in this use-case. In the meantime, you can learn more about the benefits of Edge AI here.

Application principle

Utilizing a robust two-stage detection algorithm, the system processes vehicle images to detect and recognize license plates. When a plate is successfully identified, the system displays the extracted license plate number alongside its grayscale image patch, providing a clear visual reference.

Approach

PerCV.ai's platform for on-device Vision Intelligence, builds an efficient and robust ANPR solution using computer vision and AI at the edge, running in real time and relying on a proprietary ML engine. The two-stage detection algorithm ensures high accuracy, robustness, and expandability. The system communicates with a serial API, offering integration with 3rd party software and hardware components

Sensor

RGB Image sensor.
The system includes an image sensor/camera, that is an IMX335-based, MIPI CSI that comes with the STM32N6570-DK.

Results

The system processes 3.1–6.0 FPS end-to-end, captures plates within 0.75–2m at ±10° viewing angles, and operates effectively in 300–20,000 Lux conditions.

It tracks up to 5 plates simultaneously up to 6 FPS.

Author: Irida Labs | Last update: February, 2025

Find out more

Resources

Optimized with STM32Cube.AI

A free STM32Cube expansion package, X-CUBE-AI allows developers to convert pretrained AI algorithms automatically, such as neural network and machine learning models, into optimized C code for STM32.

Most suitable for STM32N6 Series

The STM32 family of 32-bit microcontrollers based on the Arm Cortex®-M processor is designed to offer new degrees of freedom to MCU users. It offers products combining very high performance, real-time capabilities, digital signal processing, low-power / low-voltage operation, and connectivity, while maintaining full integration and ease of development.