关键词: 机器视觉, 驾驶人行为检测, Gabor小波, 特征点

Abstract: With multi-scale and multi-directional characteristics，Gabor wavelet is often used for texture analysis and feature extraction studies，and obtained more applications in the field of driver fatigue monitoring technology.First，the basic principles of Gabor wavelet and using methods are introduced，and then with examples technology segments of driver facial feature point tracking are focused on，including the method of determining Gabor wavelet parameters and the method of extracting the feature point’s eigenvector.And finally the three candidates with the maximum similarity rules are identified.For the different image conditions，reasonable proposals in the actual application of Gabor wavelet are proposed.Practice has proved，the Gabor wavelet tracking algorithm has higher accuracy.The effect is very ideal，and lays a good foundation for the driver fatigue analysis.

Key words: machine vision, driver behavior surveillance, Gabor wavelet, feature point