Access Control Through Mask Detection and Estimation of People Capacity in Covered Premises

The health emergency due to the COVID-19 pandemic requires the search for technological and intelligent solutions that facilitate the control of biosecurity measures such as social distancing, to use of a mask, and capacity in covered spaces. This work aims to develop a prototype based on artificial vision algorithms, capable of performing the automatic mask detection and people counting who go to covered premises such as bars, restaurants, gyms, cinemas, and micro-market among others. The prototype implements SSD-MobileNet object detection and SORT tracking algorithms that work on the electronic device NVIDIA Jetson Nano, equipped with two video cameras to perform mask detection and people counting respectively, as well as speakers, for emission of audible alert messages about the use of mask and the capacity estimation within the premise and an external web server too in which people counter information is displayed.