Intelligent detection of mask wearing based on cloud-edge collaboration

With the continuation of the Corona Virus Disease 2019 (COVID-19) epidemic, the epidemic prevention experience of various countries has confirmed that wearing a mask is an effective way to reduce the spread of the virus. With the repeated epidemics, China's epidemic prevention has entered a normalized management, and people's awareness of protection has declined to a certain extent compared with the initial stage of the epidemic, and there have been cases of incorrectly wearing a mask or even not wearing a mask. Some regulatory and technological means are necessary to stabilize the results of epidemic prevention and remind people to wear a mask in public. This paper proposes an intelligent detection scheme for mask wearing based on cloud-edge collaboration, which uses the You Look Only Once (YOLO) target detection algorithm to detect the situation of people wearing a mask in public places. This solution introduces the cloud-edge collaboration mechanism into the Internet of Things (IoT). Under this mechanism, the cloud center uses a larger image dataset to train the mask detection model, and the trained model will be released to the intelligent detection camera through IoT. In this way, the camera at the edge site can accurately detect the case of mask wearing. This mechanism can make full use of cloud center resources, improve the detection speed of edge site cameras, and promptly correct the behavior of incorrectly wearing a mask. It only sends the detection results to the cloud center instead of the original images, which can effectively reduce the transmission bandwidth of IoT. The experimental results show that the intelligent detection scheme of mask wearing with cloud-edge collaboration mechanism can achieve better detection accuracy.