采用ELM的基于眼部状态的驾驶员疲劳检测方法的研究

发布时间：2018-01-04 02:11

本文关键词：采用ELM的基于眼部状态的驾驶员疲劳检测方法的研究　出处：《东北大学》2014年硕士论文　论文类型：学位论文

【摘要】：目前,交通安全问题成为世界性的广泛话题,根据统计,大约有40%的交通安全问题来自疲劳驾驶。研究驾驶员疲劳监测方法,用以预防交通事故的发生,具有非常重要的现实意义。本文在对现有的驾驶员疲劳检测相关技术进行深入分析的基础上,对视频分析方法进行了深入的研究,提出了一种采用ELM(Extreme Learning Mechine,极限学习机)技术的、基于眼部状态的驾驶员疲劳监测方法。首先,通过基于皮肤颜色的视图分割方法来进行检测驾驶员的脸部。接着采用KLT(Kanade-Lucas-Tomassi)算法,对脸部检测数据进行跟踪并截取出每帧图像。然后,通过使用一个积分映射函数用来估测眼睛的区域,以检测眼睛张开或者闭合的状态。进而计算一组帧图像中的眼睛开闭程度距离,并基于上述数据进行特征抽取(眨眼频率和眼睛关闭时间)。由于在面向视频的计算中进行多次迭代计算,数据量将非常庞大而且时间复杂度很高。为了降低时间复杂度,减少计算量,引入了滑动窗口的数据处理方式进行实时计算。最后,本文采用提取出的特征数据对ELM进行训练,从而对驾驶员的疲劳状况进行检测和分类。本文利用实际的网络摄像头的实时视频数据进行了实验研究,结果表明,本文所提出的方法在保证实时性要求的前提下,能够有效地提取驾驶员的眼部特征,准确地判断驾驶员的工作状态。
[Abstract]:At present, traffic safety has become a worldwide topic. According to statistics, about 40% of the traffic safety problems come from fatigue driving. The method of driver fatigue monitoring is studied to prevent traffic accidents. It has very important practical significance. Based on the in-depth analysis of the existing driver fatigue detection technology, the video analysis method is deeply studied in this paper. This paper presents a driver fatigue monitoring method based on eye state using ELM(Extreme Learning Mechine (Ultimate Learning Machine) technology. The skin color based view segmentation method is used to detect the driver's face. Then the KLTT Kanade-Lucas-Tomassian algorithm is used. The face detection data are tracked and each frame image is captured. Then, an integral mapping function is used to estimate the eye region. To detect the state of eye opening or closing, and then calculate the distance of eye opening and closing in a set of frame images. Based on the above data, feature extraction (blink frequency and eye closure time) is carried out. The amount of data will be very large and the time complexity will be very high. In order to reduce the time complexity and reduce the amount of computation, the sliding window data processing method is introduced for real-time computation. In this paper, the extracted feature data is used to train the ELM to detect and classify the driver fatigue. This paper makes use of the real time video data of the webcam to carry on the experimental research. The results show that the proposed method can effectively extract the eye features of the driver and accurately judge the driver's working state on the premise of ensuring the real-time requirement.
【学位授予单位】：东北大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U463.6;U495;TP391.41

【相似文献】