Research Of Fatigue Driving Detection Algorithm Based On Eye Detection

With the progress of society, cars have become the most important mean of transportation. The rise in the number of cars also leads to a surge in traffic accidents, and the frequency of traffic accidents caused by fatigue driving is rising. How to monitor drivers’fatigue state in driving process and alert in time in order to effectively reduce the occurrence of traffic accidents has become a hot research topic.This paper examined current driver fatigue detection methods commonly used in domestic and abroad, compared and analysed the advantages and disadvantages of various methods. In consideration of experimental results and real-time&accuracy requirements of driver fatigue detection, this paper uses PERCLOS principle assisted by blinking frequency to judge fatigue, successfully improved PERCLOS principle.This article evaluates fatigue mainly through image-processing technology from extracting real-time information of driver’s eyes. Gradually narrow the scope of recognition by positioning the steps. Firstly locates faces based on skin color features, uses the YCbCr color space because of its apparent clustering of skin colour features for video image segmentation, finds face area through integral projection. Then determines human eyes candidate region by skin color segmentation with multiply integral projection method, pinpoint human eye area through integral projection, extracts eyes effective information for judging eye state and blink frequency. Finally uses PERCLOS-P80for fatigue detection. Some effects during driving such as strong light will stimulate human eyes, the reaction of driver will cause a significant rise of the PERCLOS parameter f, it makes fatigue test results inaccurate. This article puts forward a new method of blink frequency assisting PERCLOS-P80evaluation standard, this method effectively avoid the interference caused by special external stimulation, increases the relevance and accuracy of PERCLOS-P80standard.The method used in this article with advantages of speed, less time consuming and high accuracy. It can accurately detect driver’s real-time status in driving process, and alert fatigue in time, in order to avoid the occurrence of traffic accidents due to fatigue driving.