In light of the ongoing pandemic, nations around the world are putting into place various practices to help curtail the spread of COVID-19. One common rule among these is the requirement to wear masks or personal protective equipment (PPE) out in public, as well as barring anyone with a fever. However, observing large numbers of people in such detail can be overwhelming for security teams and staff. AAEON’s BOXER-8120AI is being deployed by developers to assist in monitoring for such breaches of newly imposed rules.
On April 3rd, the task force in charge of the Taiwan CDC’s efforts, the Central Epidemic Command Center (CECC), imposed new rules requiring the wearing of masks on public transportation, with stiff fines for failing to comply. Also included were rules restricting access to those whose body temperature is above 37.5°C (99.5°F). With an average of over two million passengers each day, monitoring and enforcing these rules on Taipei’s Metro system is daunting to say the least.
Utilizing artificial intelligence, authorities can reduce the workload of the task, allowing security and staff to effectively monitor passengers without requiring a ramping up of personnel. Some developers are deploying the BOXER-8120AI embedded edge platform to power this innovative and important role.
- Nvidia Jetson TX2
- HMP Dual Denver 2 + Quad ARM A57
- Support AAEON ACLinux 4.9, Compliance with Ubuntu 18.04
- 8GB LPDDR4, 32GB eMMC 5.1
- RS-232 x 2, USB3.0 x 2, OTG x1
- HDMI 2.0 x 1, LAN x 4 (Intel i211*3, Native*1)
- Micro-SD x 1, Remote On/Off
- Power input: 10~24V
Traditional monitoring equipment usually involves only using infrared cameras. However, the surrounding environment can interfere with readings, leading to missed reports and false alarms, with the requirement to adjust data afterwards. Utilizing AI technology with deep learning, the application can combine both facial recognition with thermal detection to accurately detect those who are violating mask requirements, or have fevers.
The BOXER-8120AI is currently deployed in several hospitals, providing monitoring of hospital entrances. In this role, it is crucial to detect potential patients, allowing hospital staff to separate and administer the appropriate treatments. Despite wearing a mask, the facial recognition models can still accurately identify people and features, allowing hospital staff to quickly locate and address anyone with abnormally high body temperature. Operating at the edge, the system is able to identify and measure the temperature of up to 240 people per minute and can allow for the monitoring of multiple entrances, greatly reducing the workload of staff.
The BOXER-8120AI is one of AAEON’s most popular AI Edge Computing solutions. It’s compact size and rugged, fanless design are perfect for use and deployment in just about any role or application. Powered by the NVIDIA Jetson TX2 SOC with 256 CUDA cores, it supports AI frameworks such as TensorFlow, Caffe2, and can support custom AI inference software. With four Ethernet ports, two USB 3.0 and two COM ports, the BOXER-8120AI provides flexible I/O to easily integrate with most projects.
“The BOXER-8120AI provides a flexible platform that can power new projects or easily integrate into existing equipment or infrastructure, whether it’s transportation, retail, factories, parking lots, schools, or wherever your project demands,” said Alex Hsueh, Senior Director of AAEON’s System Platform Division. “One of our customers have refined this application to contact relevant personnel in real-time via mobile phones or text messages; as well as provide quick interpretation after the fact to prevent future incidents. At present, this customer has completed prototyping and is set to order 1,000 units for their application,”