“Setting a new standard for license plate recognition”

Seyeon Tech is launching a 4-megapixel global shutter IP camera module (Model: FWC-EX1-567) optimized for license plate recognition. This camera module effectively addresses the challenge of simultaneous recognition of reflective and non-reflective license plates, significantly enhancing license plate recognition accuracy. Further details are provided below.

Table of Contents :

1. Difference between reflective and non-reflective license plates

2. Challenges of license plate recognition (LPR)

3. Obtaining stable license plate recognition images under various day and night conditions with our own ISP and image processing algorithm

4. Others

Outline :

1. Difference between reflective and non-reflective license plates

The 8-digit reflective license plates introduced in July 2020

License plates are categorized into reflective and non-reflective types based on how they reflect light. Traditional IP cameras faced challenges recognizing license plates of differing reflection types when configured for one specific type. For instance, setting the camera for reflective plates caused non-reflective plates to appear dark, while setting it for non-reflective plates led to reflective plates appearing overexposed and washed out. The new product successfully commercializes advanced ISP (Image Signal Processing) technology and video stream transmission technology, enabling simultaneous resolution of these issues. 

2. Challenges of license plate recognition (LPR)

Typically, license plate recognition rates decline at night, prompting IP cameras to employ various lighting technologies to address this issue. On the other hand, during the day, factors such as shadows on license plates or overexposure caused by the sun’s position can lead to inconsistent license plate recognition. 

3. Obtaining stable license plate recognition images under various day and night conditions with our own ISP and image processing algorithm

The new IP camera module is designed to account for these complex environmental factors, utilizing its built-in ISP and image processing algorithms to deliver stable license plate recognition under various day and night conditions. Furthermore, the captured footage is compressed and transmitted as a separate video stream, enabling selective recognition processing at the recognition stage.

The module is also equipped with robust core functionalities. It supports ONVIF Profile S/T, ensuring high compatibility with various NVRs and VMS platforms. Additionally, it meets TTA security certification and National Intelligence Service security standards, making it reliable for use in both public institutions and private sectors. The module offers additional features such as DI/DO, two-way audio, and optional MicroSD support. By utilizing the ‘FlexWATCH Open AI SDK,’ users can implement not only the previously mentioned LPR (License Plate Recognition) but also a variety of AI features tailored to their needs. Customers can develop AI functionalities independently using the SDK or receive support from Seyeon Tech

.

The module has completed all internal testing and is scheduled for official sale starting in January 2025. Patent applications for its features have also been finalized. Currently, it supports only the 4-megapixel global shutter module, but there are plans to apply it to rolling shutter IP camera modules as well. With the launch of this new product, it is expected to set a new standard in license plate recognition, particularly in areas where license plate recognition is crucial, such as parking management, traffic enforcement, and vehicle access control.

Others

Seyeon Tech is the oldest specialized development and export company of IP camera modules and complete products in South Korea. The company supplies global shutter modules for license plate recognition in large quantities both domestically and internationally, and it also possesses middleware optimized for various SoCs and operating systems. Recently, with the release of its open AI SDK, Se-Yeon Tech has been strengthening its collaboration with several AI software companies.