【摘要】 提出一种基于普通摄像头快速有效的人眼定位和睡意检测方法,对光照、姿势和背景都具有较好的鲁棒性。首先从USB摄像头截取头肩部图像,经过光照补偿、运动检测和肤色处理粗定位人脸区域,再根据眼睛亮度、色度信息,数学形态学处理将眼部特征最大化,最后使用模糊逻辑筛选候选眼睛块,Hough圆检测判定眼睛状态。更多还原

【Abstract】 A fast eye location and drowsiness detection method with robustness in various illumination,gestures and intricate backgrounds is presented.Firstly,the images of head and shoulder are acquired from common USB camera.Then,processes of illumination compensation,motion detection and skin color segmentation are carried out,and a rough face area is obtained.The eye features are extracted based on eye luminance and chroma with the use of mathematical morphologic processing.Finally,the candidate eye blocks are filtered by fuzzy logic,and the eye state is detected and judged by Hough circle.更多还原