基于视频的驾驶疲劳检测技术的研究  

Researching Driver Fatigue Detection Using Video Technology

作　　者：邓正宏[1] 黄一杰[1] 李翔 张天凡 

机构地区：[1]西北工业大学自动化学院,陕西西安710072 [2]西安特种设备检验检测院,陕西西安710068

出　　处：《西北工业大学学报》2015年第6期1001-1006,共6页Journal of Northwestern Polytechnical University

基　　金：国家自然科学基金(F011102)资助

摘　　要：疲劳驾驶已经成为交通事故的重要因素,若能及时监测驾驶员疲劳程度并且对其进行警告,则可降低此类交通事故的发生率。在图像处理的基础上从驾驶员实际状况出发,从背景中分离驾驶员面部区域,分别采用优化等照度线法和优化mouthmap法提取眼睛和嘴巴特征参数,先在模糊神经网络的基础上建立疲劳分类器识别驾驶员疲劳程度,再在DSP系统上去实现疲劳驾驶检测系统。实验结果表明,该系统满足了一般疲劳的动态识别要求,具有较强的实用性。Fatigue driving has become an important factor in traffic accidents; warning through timely monitoring of driver fatigue may reduce the incidence of traffic accidents. Using the image processing and the driver＇s actual situ- ation, we separate driver＇s face in the background region, use optimized equal illumination method and optimized mouthmap method respectively to extract characteristic parameters of the eyes and the mouth, firstly establish fatigue classifier to identify driver fatigue based on fuzzy neural network classifier, and then implement the driver fa- tigue detection system in DSP system. Experimental results and their analysis show preliminarily that the system with strong practicability can meet the requirements of dynamic recognition of general fatigue.

关 键 词：疲劳驾驶 面部特征 面部识别 眼睛监测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]