Remote Surveillance System for Driver Drowsiness in Real-time Using Low-cost Embedded Platform

This paper present a non-intrusive computer vision-based driver drowsiness surveillance system. It based on a system design for commercialize, which adopted active IR illuminators and related hardware/software implementation to acquire the drowsy-related visual cues and defected the driver fatigue. The function of driver fatigue detection has been modularized on low-cost DSP platform. Through the recording and 3G transmission module, the signals include: level of driver fatigue, real-time in-lout-cabin video stream and vehicle location, etc. will be monitored by remote control center in real-time. The system has been adopted by the proving ground of Automotive Research & Testing Center (ARTC, Taiwan) to monitoring their testing-drivers in their testing drive.