基于改进ASM的多特征融合疲劳检测方法  

Driver’s fatigue detection based on improved active shape model and fusion of multi-clues

作　　者：陈鑫 李为相 李为 张文卿 朱元 CHEN Xin;LI Wei-xiang;LI Wei;ZHANG Wen-qing;ZHU Yuan(College of Electrical Engineering and Control Science,Nanjing Tech University,Nanjing 211816,China)

机构地区：[1]南京工业大学电气工程与控制科学学院

出　　处：《计算机工程与设计》2019年第11期3269-3275,共7页Computer Engineering and Design

基　　金：江苏省"六大人才高峰"基金项目(XXR-012)

摘　　要：为解决驾驶员在行驶过程中头部发生多角度变化导致难以定位面部特征的问题,应用改进的ASM算法精确定位眼睛和嘴部区域,计算眼睛的长宽比值、嘴部高度值和嘴部附近的黑白像素比值,得出眨眼频率和嘴巴张开程度,将眼部状态和嘴巴的张开程度作为模糊推理机的输入,得出三类疲劳水平,准确量化疲劳程度。实验结果表明,该非入侵式疲劳驾驶检测方法将经典ASM算法分类能力的结构误差降到了最小,该模糊推理系统对检测驾驶员疲劳程度和提高行车安全性方面是有效的。To address the problem that it is difficult to locate the facial features when the driver’s head changes during the driving process,the improved ASM algorithm was proposed to accurately locate the eye and mouth area,and the aspect ratio of the eye,the mouth height,and the black-and-white pixel ratio near the mouth were calculated to obtain the blink frequency and the degree of mouth opening.The state of the eye and the degree of mouth opening were used as the inputs of the FIS to obtain the three types of fatigue levels,so as to accurately quantify the degree of fatigue.Experimental results show that the non-invasive fatigue detection method reduces the structural error of the classical ASM algorithm classification ability to the minimum.The fuzzy inference system is effective for detecting driver fatigue and improving the safety of driving.

关 键 词：疲劳检测 主动形状模型 眨眼检测 打哈欠检测 模糊推理系统 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]