Research On Fatigue Driving Detection Algorithm Based On Support Vector Machine

Since the reform and opening up,China's economy has begun to develop rapidly.With the economic development,people's quality of life is getting higher and higher,and cars have gradually begun to enter every household and gradually become people's production and life.An indispensable and important tool,which not only greatly improves the efficiency of people's work and travel,but also adds a lot of color to people's material life.Especially in recent years,with the continuous development of the economy,the number of cars in our country has started to soar.At the same time,there have been some negative social problems,including more and more traffic accidents.The causes of traffic accidents are,Drivers are drunk driving,driving in violation of traffic rules,emotional driving and fatigue driving,etc.According to the statistics and analysis of the causes of traffic accidents for many years,fatigue driving is an important factor causing traffic accidents.Therefore,in order to protect drivers for the safety of life and property,it is particularly important to detect fatigue driving of the driver.By carefully analyzing the research status of fatigue driving detection at home and abroad,this paper proposes a fatigue driving detection method based on support vector machines,which combines eye state and head posture,and uses improved support vectors with better classification performance.Machine to judge fatigue driving.The main work completed in the paper is as follows:(1)First analyze skin color information and commonly used color space,select the Ycbcr color space with better skin color clustering to separate the skin color part in the image to be detected,and then separate the skin color part area using Harr-like feature AdaBoost classification algorithm Perform face detection.(2)Eye detection and eye state feature extraction are performed on the detected face area.In this paper,the application of gray integral projection method and Hough transform circle detection algorithm in human eye detection is studied in detail.The detection algorithm combining the integral projection and the Hough transform circle detection locates the eyes,that is,firstly use the gray integral projection algorithm to frame the approximate location area of the eye,and then use the Hough transform circle detection to accurately locate,and then detect the eye opening Compared with the traditional eye positioning algorithm,this algorithm has a higher accuracy rate.The video collected by the camera is processed by frame processing,and the features of the eye opening and closing state feature extraction and head are performed on each frame of the image.Feature extraction of the external pose.(3)The fuzzy system and support vector machine are combined and applied to the determination of fatigue driving.The eye closing rate,the number of blinks,and the degree of eye closure per unit time are used as the eye feature parameters.The head is relative to the x-axis and y-axis.The z-axis deflection angle is used as the head posture characteristic parameter in combination with the improved support vector machine for fatigue driving judgment.Compared with the result of fatigue driving judgment based on the traditional support vector machine,this method has higher accuracy.