The drive fatigue pattern monitor for haul truck drivers in surface mining operations

Abstract

Haul trucks are one of the most important transportations in the operations of large surface mines. The traffic accidents involved haul trucks result in serious damages on the vehicles, even the loss of the drivers' lives. Drive fatigue is one of the key features that cause traffic accidents in surface mining operations. The light varies and driver's behaviors in the cabin increase the difficulties of drive fatigue monitoring. This paper developed the haul truck driver fatigue monitoring prealarm methodologies using real-time, non-intrusive machine vision technologies. It detected and tracked the driver's face by using AdaBoost algorithm. It extracted the key features of driver's eyes by template matching algorithm. This paper improved the SNAKE algorithms in the monitored image processing and pattern recognition, which increased the efficiency of the fatigue detection under the light variations and different driver's gestures. It observed the eye's closure duration and calculated the cumulative total percentage of a particular time (PERCLOS) to determine the driver's fatigue status. This research was applied to a driver of the haul truck in a coal mine. The detection results indicate that it can identify the driver's fatigue. The further research of the drive fatigue monitoring will increase the reliability and reduce uncertainty in surface mining operations.