摘要
本系统采用STM32F103为无接触测温板块控制核心,搭配mlx90614测温模块,oled显示,来完成对人体和物体的分类测温,k210用做身份识别模块处理核心利用其搭载的双64核处理器以及kpu神经网络处理器,对采集数据进行神经卷积运算,身份算法方面使用了mtcnn特征点采集算法,Haar人像分类器以及196维人脸比对模型来实现高精度人脸识别,口罩识别模块我们使用了自己训练的yolo2分类模型来实现口罩分类功能。
This system adopts the STM32F103 as control core,the contactless temperature measurement plate collocation mlx90614 temperature measurement module,oled displays data to finish the categorization of the human body and the object temperature measurement,k210 used as identification module processing core use with a dual core processor and 64 kpu neural network processor to collect data for nerve convolution operation,identification algorithm use the MTCNN feature point acquisition algorithm,like Haar classifier and 196—d face matching model to achieve high precision face recognition,For mask recognition module,we used the yolo2 classification model trained by ourselves to achieve the function of mask classification.
作者
任宇逍
詹毅
张飞周
REN Yu-xiao;ZHAN Yi;ZHANG Fei-zhou(College of Optoelectronics and Information Engineering,Fujian Normal University,Fuzhou 350117,China)
出处
《价值工程》
2021年第25期176-178,共3页
Value Engineering
