A Pattern Recognition System About Eyeball Changes Based On Computer Image Processing

With the development of computer technology,computer image processing and pattern recognition play a more and more important part in modern times,and the Pattern Recognition System belongs to application of the technology.A same problem all countries are facing is traffic safety.Many taffic accidents are caused by tire of drivers,especially in China,for the drivers in China work longer than other countries'drivers.According to information provided by traffic police department,we know that traffic accidents which caused by drivers'tire is 40 pecent of all traffic accidents. Using computer image processing technology and pattern recognition technology,we can recognize whether the driver is tired or not.When the driver is tired,he should be reminded or forced to have a test.According to images including eyeball,basic algorithms about computer image processing and pattern recognition are discussed.The algorithms are image point operators,image geometry operators,image enhancement,image segmentation, clustering and Support Vector Machine.Firstly,using color characteristics and geometric features of eyeball image,get eyeball image from original image recuring to some alogrithms including image format conversion,image graying,threshold segmentation,edge detection and extraction.Secondly,getting data from eyeball image using feature selection and extraction algorithm.Thirdly,training data by support vector machine and create classifier.At last,recognizing whether a driver is tired or not.Based on the domestic and international pattern recognition research about eyeball changes,using the technology of computer image processing and pattern recognition, the paper analyzes image clustering segmentation,pattern feature selection and extraction, and training data by support vector machine.Experiments proved that the system can accurately identify whether a driver is in a state of tire or not.