ABSTRACT
The COVID-19 pandemic has affected a variety of aspects of our everyday life. Most activities like entertainment, healthcare, education and businesses have been reshaped due to the safety guidelines. Proper monitoring in indoor areas is essential to limit the spread of COVID-19. This paper presents a low-cost prototype system that addresses the indoor safety issue by combining a mask detector and temperature measurement system with a smart wearable band which alerts people to maintain social distance in close vicinity. The focus is on ensuring safe distance, wearing a mask, and no entry for people with high temperatures. Firstly, the mask and temperature system has an Arduino NANO that works as the primary device. The Arduino is connected with an ESP32-Cam that sends the image to a client where we have trained and developed a machine learning model using thousands of masked and unmasked pictures. Following, the model uses an image classification algorithm with the tensorflow.js model and gives us the result with an accuracy percentage. Secondly, the temperature is measured with the help of an MLX90614 non-contact sensor. The temperature of a person is also shown on the monitor at of. Finally, a wearable device is presented with a NodeMCU 8266 Wi-Fi module. It uses Received Signal Strength Indicator (RSSI) value to detect another similar device and alerts through a vibrator and buzzer if the social distance rules are violated. We evaluated the system in real-life scenarios, and the mask detection system gives an average accuracy of 98.7%. We have presented an in-depth analysis of the Mask Detection System, showing different mask types, the accuracy of the machine learning algorithm, temperature measurements and results. Similarly, the distance measurement system is presented with several factors. © 2022 IEEE.