基于人脸跟踪云台的红外热成像系统  

Infrared Thermal Imaging System Based On Face Tracking PTZ

作　　者：张紫东 马春燕[1] 陈燕[1] 姚坤 李根 郭然 ZHANG Zidong;MA Chunyan;CHEN Yan;YAO Kun;LI Gen;GUO Ran(College of Electrical and Power Engineering,Taiyuan University of Technology,Taiyuan Shanxi 030024,China)

机构地区：[1]太原理工大学电气与动力工程学院,山西太原030024

出　　处：《电子器件》2022年第3期665-671,共7页Chinese Journal of Electron Devices

基　　金：山西省回国留学人员科研资助项目(2020-039)。

摘　　要：针对疫情防控与疫情常态期间的公共场所人员温度异常检测模式固定、响应迟钝、直接接触测温等缺点,设计了一种追踪动态人脸的红外热成像测温系统。系统基于红外阵列传感器MLX90640和CMOS摄像头模块OV7725,以STM32H743为核心控制器,加持了二自由度云台,并借助OpenMV IDE软件结合HaarCascade级联分类器、目标追踪PID算法、LBP特征提取,实现了从人脸检测、人脸跟踪到温度检测和人脸识别的功能。最后通过整机调试,可以完成对动态人脸的追踪、人体温度异常的检测、被测者身份认证。实验表明,该非接触测温系统操作简单、红外热成像图像质量较高、测温灵敏,且硬件便于与门禁系统搭配使用,应用场景较广。Aiming at the disadvantages of fixed temperature anomaly detection mode,slow response and contact temperature measurement in public places during the prevention and control period and the normal period of the epidemic,an infrared thermal imaging temperature measurement system is designed to track the dynamic face.The system is based on the infrared array sensor MLX90640 and CMOS camera module OV7725.The system takes STM32H743 as the core controller and has two-degree-of-freedom PTZ.Such functions as face detection,face tracking,face recognition and non-contact temperature detection are realized through the combination of the HaarCascade cascade classifier,target tracking PID algorithm and LBP feature extraction by using the OpenMV IDE software.Finally,after the debugging of the whole system,the tracking of dynamic face,the non-contact detection of human body temperature anomaly,and the identity authentication of the tested are fultilled.The experiment shows that the non-contact temperature measurement system has the advantages of simple operation,high image quality,sensitive temperature measurement,and easy to be integrated with the access control system.

关 键 词：温度监测 MLX90640红外阵列传感器 OpenMV 目标追踪 人脸识别 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]