Face Mask Detection using Haar Cascade Classifier Algorithm based on Internet of Things with Telegram Bot Notification

Abstract

The lack of public awareness of wearing masks during the COVID-19 pandemic is one of the causes of the high number of Covid-19 cases in Indonesia. Since the beginning of June 2020, the government has set a New Normal phase. This is done to restore the economy and prevent the spread of the COVID-19 pandemic. During New Normal, activities can still run by implementing health protocols by requiring masks to be worn. Currently, the detection of masks is still done manually by security officers because of the fatigue factor, so that human errors often occur. To overcome this, an automatic system is needed to detect people wearing masks and not wearing masks. In this study, a mask detection system was made using the haar cascade classifier method by utilizing machine learning, image processing, and the internet to facilitate connectivity. The result of this research is an internet of things-based mask detection system using the haar cascade classifier method that runs on a raspberry pi to monitor and distinguish between people with masks and not masks in various light conditions with the help of an additional IR (Infrared) module on the camera. If a person is detected who is not wearing a mask, the program will automatically capture it, and an alarm will sound and send the captured results to the telegram bot. The resulting performance is when the video stream reaches 12-60 fps, the system can run well without stuttering even during the video stream. The connection speed to the telegram bot got excellent results without any delay with an average time of 0.001695977 seconds with a maximum detection distance of 1.2 meters.