Implementación de un prototipo de detección de mascarrilla y medición de temperatura para el personal que ingresa a la empresa PAUFIT, usando técnicas de visión artificial.

Abstract

This work describes the design and implementation of a mask detection prototype and body temperature measurement, using artificial vision techniques, as a control device for personnel entering the PAUFIT company. With the learning algorithm performed on the TensorFlow platform and the artificial vision algorithms developed on the Python platform making use of the tools with the OpenCV library, it was developed the algorithm for the recognition and control of the use of face masks, necessary in the pandemic that has been given globally, to this was linked a protocol of alert signals managed by an Arduino Uno module that detects the presence and measures the user's body temperature; additionally, it linked a control signal to play a voice message, which informs the people if they meet the requirements of a mask and body temperature, thus managing admission personnel using the MQTT communication protocol. The structure of the prototype was designed in SolidWorks giving an aesthetic and suitable finish for the project. Through tests of equipment positioning through image acquisition, an efficiency of 92% was verified in mask recognition when the person is 100 cm from the sense station. For the evaluation of the prototype, real-time tests were carried out in 4 different scenarios, thus determining a device efficiency of 94.24%. In conclusion, the device meets the objectives set.