基于人脸识别的车辆安全系统  

Vehicle Safety System Based on Face Recognition

作　　者：黄原有 阳韬 张杨志 Huang Yuanyou;Yang Tao;Zhang Yangzhi(Guangzhou College,South China University of Technology,Guangzhou 510800,China)

机构地区：[1]华南理工大学广州学院,广州510800

出　　处：《机电工程技术》2021年第5期112-114,共3页Mechanical & Electrical Engineering Technology

基　　金：华南理工大学广州学院大学生创新创业训练计划项目(编号:56JY200508)。

摘　　要：研究运用卷积神经网络实现对司机的人脸识别,以提高车辆的安全性能。系统事先采集车辆司机的照片,进行模型的训练。在应用当中,通过摄像头采集驾驶位的驾驶人员图片,并运用树莓派上传到云服务器端,将采集到的照片导入已训练的模型,进行图片的识别。如若识别图片为司机,则开启汽车电源,否则将发送消息给车主,并发出警报和关闭汽车电池。经过测试,模型的识别率大于80%,因此可运用人脸识别技术来提高车辆的安全性。In order to improve the safety of the vehicle,convolution neural network was used to realize the driver's face recognition.The system collected photos of the vehicle driver in advance to train the model.In the application,the driver's picture in the driving position was collected through the camera,and uploaded to the cloud server using the Raspberry Pi,and the collected photos were imported into the trained model for picture recognition.If the picture was identified as the driver,turned on the power of the car,otherwise,a message would be sent to the owner,an alarm would be issued and the car battery would be turned off.After testing,the model's recognition rate is greater than 80%,so face recognition technology can be used to improve vehicle safety.

关 键 词：人脸识别 车辆安全 卷积神经网络 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]