In China, many ticket gates at transportation stations are being fitted with the newest facial recognition technology to heighten security and improve social control. Traditionally, such technology was mainly deployed at international airports as part of security checks against terrorists or criminal threats. However, as the technology has become more sophisticated and available with advancements in cloud technology and artificial intelligence, images can be analysed in real time with much higher speed and accuracy. China is now expanding the use of facial recognition to entry points of railway stations and other large passenger transportation stations.
Our customer in this project wanted to incorporate facial recognition technology into their ticket gate application, which uses USB cameras or network (LAN) cameras to capture the facial image of anyone attempting to gain entry. The system will then run the image through a database of mug shots of suspected criminals or blacklisted persons to find if there is any match. Meanwhile, a card reader or ID certificate reader will also check out personal information of the passenger and compare it against a database to help identify the person. If the system does not detect any suspicious circumstances, the person will be allowed to enter the station through the gate—otherwise an alarm will sound, LED indicators will flash, and the person could end up being detained.
Application Requirements
The system had to be able to quickly capture real-time images and perform recognition, verification, or identification within a very short period of time to ensure the smooth flow of people through the station. It required a high performance computing platform (Intel Core i7 grade or above) to quickly run complex algorithms and facial recognition software, and a good network connection was required to link all the gates to a centralized management platform and an image database. Also, because the computer was to be embedded into each ticket gate and had to connect to a number of devices, a computer with a small physical footprint and multiple I/O was essential.
Solution
Though high computing performance requirements usually run contrary to small device footprints, Advantech’s ARK-3520L fanless box computer met both requirements with a unique mechanical and ventilation design, delivering Intel 6th generation computing capability and up to 32GB memory support in a 220 x 101 x 233 mm small form factor.
ARK-3520L also features multiple I/O ports supporting serial communication, USB cameras, and network (LAN) cameras, making its adaptable to many application scenarios with varied space constraints and network support. High flexibility was especially important in this case as the embedded computer was not only used in the new ticket gates but also used for retrofitting to existing gates to upgrade them in order to enable facial recognition.
In terms of software support, ARK-3520L was deployed with WISE-PaaS/RMM software enabling remote device monitoring and management services with an Advantech-prepared application-ready IoT/cloud platform. The customer could also choose to use Microsoft Azure or Amazon Web Services as their cloud service platform.
Applications such as this usually come with concerns over data security, so the customer also deployed Advantech’s WISE-PaaS/Security platform for centralized security management to protect the system against cyber hacks and to safeguard system settings and data integrity if any failures occur.
Benefits
- High computing power and memory for cutting-edge facial recognition applications
- Multiple I/O support delivers flexibility and expandability
- Small footprint for limited-space installations in ticket gates
- Pre-installed IoT software for easy deployment of remote monitoring functions
- Optionally purchased security software platform provides comprehensive data and system protection
For more information, click here.
-205x205.jpg)
