DETECÇÃO DE FADIGA A PARTIR DA ANÁLISE DE IMAGENS FACIAIS

A large number of accidents and injuries caused by the presence of fatigue on people has caused a concern about this, more attention has been taken in recent years. Studying and developing techniques capable of detecting fatigue in a user has become possible thanks to the continuous evolution of technology and computer vision. Image processing has become a strong tool because does not interfere the driving of the vehicle, however, there are interferences that make difficult the analysis of the driver through the computer vision, these interferences are difficult to control because they involve the luminosity of the environment, cost of computational power of the tool and unnecessary objects in the environment. computer vision techniques were used: Template Matching, Hough Transfor and Landmarks, Python language with the help of the OpenCV library and use of low cost hardware such as Raspberry. The results were satisfactory and show that the combination of techniques and controlled light makes it possible to detect fatigue and alert the driver with great accuracy.