Implementación de un sistema de detección de somnolencia para conductores de vehículos mediante visión artificial

Abstract

The objective of this degree work is to apply the concepts and theories of computer vision and deep learning to road safety through the development of a driver drowsiness detection system. For the system implementation, a Raspberry Pi was used along with a camera to capture real- time images. Image processing was carried out using facial feature detection algorithms and eye state analysis to identify fatigue signs such as prolonged eyelid closure. It is worth noting that the system includes an alert module that emits acoustic signals when critical drowsiness states are detected in the driver. The implementation was designed to be energy-efficient and adaptable to different driving environments. The work concluded with the experimental validation of the system, evaluating its performance under various lighting conditions and facial characteristics.