Detecção de pessoas em imagens, implementando técnicas de visão computacional em um Raspberry Pi

Abstract

Nowadays, there's an increasing number of electronic devices present in our lives. With technological advancement of these devices, it's common that many of them remain connected to the power grid in stand-by, state in which continue to consume low amounts of energy, which together, due the large number of devices, end up generating a big waste. It is also increasing the number of environments monitored by cameras. The observation of these two facts motivated the development of this project, resulting in a system capable of detect the presence of people in images. The system was developed in C ++, implementing, with the help of OpenCV library, computer vision techniques in the Raspberry Pi platform, with the goal of, hereafter, process images acquired by surveillance cameras and control the power distribution of the monitored environment, turning it off for devices whose function is not required in the absence of people, as well as rebind it, case the presence of a person be detected.