Intelligent Body Temperature Detection and Analysis System

MEGVII Ming Ji Intelligent Body Temperature Detection and Analysis System

Ming Ji – Mini solution integrates face detection and dual sensing via infrared cameras and visible light, which is powered by our new generation AI productivity platform Brain++. The system is designed to help staff working at locations with low-to-medium foot traffic to swiftly identify people who may have elevated body temperatures, and locate them in the crowd using available facial information (please note this is not to be confused with personal identity verification). MEGVII’s proprietary light-weight, low-power-consumption, high-performance CNN (convolutional neural network) algorithm model – ShuffleNet has the capability to empower high accuracy AI algorithm model on edge analytic devices. The system can be easily setup within approximately 30 min. It allows staff to complete all body temperature screenings without close physical contact, while reducing transmission risk, enhancing personal safety, and improving passenger traffic efficiency in public places

  • “AI+Blackbody” Measurement System with Very High Cost Effectiveness Proprietary CNN on edge devices ShuffleNet to bring out best performance on edge devices
  • Dynamic Temperature Calibration Intelligent passive blackbody technology, real-time sensing of ambient temperature, and dynamically compensates for differences to cope with complex scenes
  • Temperature Measurement with High Speed Accurate temperature measurement for 3~4 people at the same time
  • Rapid Deployment Installed on edge device, 30-minute rapid deployment
  • Optimized Detection Model Optimized forehead detection model for detecting the face of people who wear hats and masks, face detection rate >95%
  • Ultra-precision Temperature Measurement Tested by the Beijing Institute of Metrology, the accuracy within the range of 30℃ ~ 45℃ is as high as ± 0.3℃ (in laboratory environment).
  • Non-contact Temperature Measurement Support up to 4 meters in non-contact temperature measurement, triggering alert when an individual with a high temperature passes through the system.
  • Mask-wearing Detection Automatically identify and flag for those who do not wear masks, detection rate >95%

Application Scenarios


  • School
  • Hospital
  • Community
  • Government Building
  • Airport Security
  • Train Station
  • Bus Stattion
  • Subway

Ref. ,