Desenvolvimento de sistema de monitoramento de câmeras para detecção de aglomeração e máscaras de proteção individual através de redes neurais

Abstract

Due to the health crisis of the new coronavirus, Covid-19, the world is faced with a public health situation where care such as the use of personal protective masks and social distancing to avoid agglomerations were necessary to prevent the spread of the virus. However, these precautions and precautions are not properly complied with by part of the population and, because of this, the need arose to create monitoring mechanisms in order to comply with public health determinations and alert in cases of non-compliance with them. Therefore, a real-time camera monitoring and management system was developed that applies digital image processing and computer vision algorithms through artificial neural networks to detect the use of masks and crowding of people. The system was developed in Python language together with the PyQt framework for the development of the graphical interface. The system was built under an object-oriented paradigm to manage and connect the backend and the frontend. In addition, a MySql database was created with 4 tables: one for storing access credentials, one for storing used camera information, one for storing non-mask infractions and one for storing agglomeration infractions.