Driver performance tracking system

In this study, an image processing based system that follows the drive performance has been developed in order to minimize the accidents caused by the drive performance. The developed Driver Performance Monitoring System is operated with an algorithm that detects and monitors driver sleep state, carelessness and driving time. While the use of technology is increasing in the automotive industry to reduce traffic accidents, one of the most widely used technologies is computer vision systems. Although various methods have been developed to prevent accidents due to driver performance, computer vision systems are preferred due to their non-invasive properties. In the proposed system, the eye, mouth and head movements of the driver are monitored with a camera without interfering with the driver. Eye closure and yawning were followed by the determination of drowsiness / fatigue of the driver, carelessness detection by following the head movements, and face recognition by face recognition. PERCLOS (Percent of eye closure) parameter, which is widely used in the literature, was used for the determination of sleepiness / fatigue from eye closure. However, PERYAWN (Percent of Yawn) parameter is used for the first time in this study. Similarly, in order to detect carelessness, PERCAR (Percent of Careless) parameter was used for the first time in this study. In addition, in this study, unlike other studies, the driver travel time was determined by PCA based driver change detection. Thus, a general Driver Performance Monitoring System (SPTS) was obtained for the monitoring of drive performance, in which the parameters used in different studies were used together. With this proposed general SPTS system, it is foreseen that it can significantly reduce accidents caused by driver performance on highways. The SPTS design developed as a result of this study can be applied primarily to all driver vehicles, as well as any work environment similar to the driver environment, for performance extraction.2020, 80 PAGE