摘要
针对目前温度测量装置的缺点,以STM32F407单片机为主控芯片,设计了一款无接触温度测量与身份识别装置.该装置通过无接触体温测量模块和身份识别模块分别提取目标温度和人脸图片,再利用基于遗传算法的BP神经网络算法、卷积神经网络和联合稀疏表示分别对读取的温度和人脸图片进行处理,进而实现温度测量和身份识别,一旦温度有异常则自动报警并将相关数据上传至后台.经测试,该装置能够对被测人进行身份识别.此外,当距离被测物1~4 cm时,测量误差在2℃以内.
Aiming at the shortcomings of the current temperature measurement device,this paper designed a contactless temperature measurement and identification device with STM32F407 microcontroller as the main control chip.The device relies on the contactless temperature measurement module and identity recognition module to extract the target temperature and facial images.Then,it uses BP neural network algorithm,convolution neural network and joint sparse representation to process the temperature and face images respectively to achieve temperature measurement and identity recognition.Once the temperature is abnormal,it will automatically alarm and upload relevant data to the background.The test shows that the device can realize identity recognition.When the distance from the tested object is 1~4 cm,the measurement error is within 2℃.
作者
邱意敏
李炜
QIU Yimin;LI Wei(School of Electrical Engineering, Anhui Polytechnic University, Wuhu, Anhui 241000, China;Anhui Key Laboratory of Detection Technology and Energy Saving Devices, Anhui Polytechnic University, Wuhu, Anhui 241000, China)
出处
《平顶山学院学报》
2022年第2期25-31,共7页
Journal of Pingdingshan University
基金
安徽工程大学国家自然科学基金预研项目(2017yyzr01)
安徽工程大学检测技术与节能装置安徽省重点实验室开放研究基金资助项目(2017070503B026-A04)
安徽工程大学青年基金项目(KZ00315012)
安徽工程大学校级科研项目(Xjky2020031)。
关键词
温度测量
人脸识别
卷积神经网络
联合稀疏表示
temperature measurement
face recognition
convolution neural network
joint sparse representation