The COVID-19 virus outbreak is changing the way we commute. Train terminals, airports, metro stations are now the virus hot zone. To limit the spread of the virus and protect passenger safety, it is required to implement the new safety measures that would maintain efficient passenger traffic flow, while automatically measuring the temperature of every passengers at the checkpoint.
AI-based thermal detection systems is introduced to deal with the passenger thermal screening. The system eliminates physical contact, processes passengers’ imaging data fast and precise, and does not require any additional workers. A AI-based thermal screening solution is consist of a thermal camera, a monitor and an edge computer for compute-intensive video analytics
LEC-2290: Compute-intensive Thermal Imaging Analytics Computer
LEC-2290 is to capture infrared images of passengers and screen for high fevers. Lanner’s LEC-2290 box is an embedded computer specially built for intelligent edge computing applications like compute-intensive video analytics. The system’s computing capacity can be expanded by GPUs to do video processing and its CPU to execute computer vision algorithms.
LEC-2290 consists of the following hardware features that allows the device to perform thermal analysis of large sets of imaging data at high speeds:
- GPU and expansion:
Includes Intel® UHD Graphics 630 to provide high graphics performance when matched with Intel Core i7 processors. LEC-2290 also supports GPU expansion.
- Powerful Intel CPU
Based on Intel Core™ i7-8700T/i7-8700, for higher energy efficiency and to run applications faster. LEC-2290’ architecture of CPU and GPU decreases the application execution time and thermal imaging output.
- I/O Connectivity for device consolidation
LEC-2290 features versatile I/O settings to consolidate various devices, like 4x USB and 4x PoE for multiple thermal cameras, 2x LAN and LTE module for wired/wireless connectivity, 2x HDMI for dual monitors outputs.
- Availability of fast memory and storage
LEC-2290 can support up to 32GB memory capacity with 2x DDR4 2400 SO-DIMM slots. These allow the compute units (CPU/GPU) to change quickly between thermal image processing to execution. Fast memory/storage allows quick data transfers so the entire thermal detection system can be faster.
- Support for Intel’s OpenVINO toolkit and the Movidius inferencing engine
OpenVINO is Intel’s toolkit that allows quick development of applications that emulate the human eye. The Movidius is a Vision Processing Unit (VPU) with deep learning capabilities that can be used to implement AI at the edge.
The AI-based temperature monitoring solution not only allows transportation authority to get the most accurate results for a massive flow of moving passengers, but provides safer way of measuring body temperature without human intervention.