驾驶疲劳监测系统DDDS设计方法  

DDDS design for drowsy driver detection system

作　　者：刘志强[1] 秦洪懋[1] 汪旸[1] 汪澎[1] 张海水[1] 

机构地区：[1]江苏大学汽车与交通工程学院,江苏镇江212013

出　　处：《江苏大学学报（自然科学版）》2008年第1期25-28,共4页Journal of Jiangsu University：Natural Science Edition

基　　金：江苏省科技厅2005年高新项目(BG2005028)

摘　　要：通过对各种疲劳监测方法进行比较分析,借助模拟驾驶仪器的研究成果,得出PERCLOS算法最适合评价驾驶疲劳程度.同时,提出一套基于PERCLOS和DSP处理器的驾驶疲劳监测系统总体设计方案,并设计开发出驾驶疲劳监测系统DDDS.监测系统应用图像差分的方法,采用灰度处理、膨胀、腐蚀等图像分析手段,实现准确定位并识别出驾驶员眼睛的睁开和闭合过程,通过分析眼睛闭合时间来判断疲劳程度.试验表明,DDDS系统对驾驶疲劳的检测准确率达到86.89%.According to the comparisons among those commonly used detecting and evaluating techniques for motor driver fatigue, the PERCLOS （percentage of eyelid closure over the pupil over time） is the most closely correlated with driver fatigue in the study of simulating driving instrument. Meanwhile, a feasible driver fatigue monitoring system design based on PERCLOS and DSP （digital signal processing） is presented and the drowsy driver detection system DDDS is developed. The methods of image difference, grey processing, erosion, dilation and so on are all used to locate the driver＇s eyes in the image and to detect the eyelid closure over the pupil over time to measure the driver fatigue degree. The results show that this scheme＇s veracity to detect driver fatigue is 86.89%.

关 键 词：驾驶疲劳 图像识别 图像处理 实时监测 

分 类 号：U461.91[机械工程—车辆工程] TP242.6[交通运输工程—载运工具运用工程]