Research on Mask Recognition Algorithm under Intelligent Monitoring

Inhalt:
In recent years, security protection is a new territory for people to respond to emergencies. Due to the appearance of some infectious diseases, it is quite important for people to wear a mask in those densely populated public places. This paper aims to study the intelligent monitoring technology of human safety protection, and adopts two sets of schemes. The first scheme is the intelligent monitoring technology based on Haar algorithm, which uses background difference method to detect moving targets. It then adopts Haar-like cascade classifier for face recognition and feature extraction, such as eyes, mouth and nose of human face, so as to judge whether people wear masks by observing whether those features exist. The second scheme is a kind of mask recognition algorithm based on convolution neural network, which consists largely of Yolov5 algorithm and makes mast recognition through training mask recognition model by Yolov5 algorithm. The experimental results show that compared with Haar algorithm, Yolov5 algorithm has higher accuracy, but the recognition speed of Haar algorithm is faster than Yolov5 algorithm.