驾驶员危险驾驶行为辨识模式研究

项目名称： 驾驶员危险驾驶行为辨识模式研究

项目编号： No.51278062

项目类型： 面上项目

立项/批准年度： 2013

项目学科： 建筑科学

项目作者： 魏朗

作者单位： 长安大学

项目金额： 80万元

中文摘要： 本申请项目综合运用人机工程学、计算机视觉技术中的嵌入式CCD图像传感采集技术、嵌入式高速DSP图像处理技术、地物形状边缘检测技术、高速运动物体形状检测分析技术、多传感器信息融合技术和模式识别技术，通过大量的交通事故现场勘察、事故案例分析、驾驶员应急操作行为实车试验、驾驶员行为测试和心理评价试验、理论分析，建立驾驶员避险操作能力及车辆响应力度计算分析模型，公路车辆运行环境（道路、标志、标线、前车状况）数据图像实时识别计算模型，驾驶员操纵车辆行驶路线图像辨识、行驶状况检测和运动轨迹预测分析计算模型，研究驾驶员在操纵车辆行驶过程中，驾驶员危险驾驶行为的界定模式和辨识模式。本课题的实施可以提高道路交通运输的安全性，有效管理和规范公路车辆驾驶员的驾驶操作行为，切实提高公路交通安全管理的科技含量与能力，为全面提高驾驶员的驾驶操作安全性和汽车主动安全性提供理论基础和技术支持。

中文关键词： 驾驶员；危险驾驶；驾驶行为；机器视觉；辨识模式

英文摘要： This project applied makes comprehensive use of ergonomics, embedded CCD image sensor collection method in computer vision technology, embedded high-speed DSP image processing technology, shape and edge detection technology for culture, shape detection and analysis technology for object in high-speed motion, multi-sensor information fusion technology and mode recognition technology. Based on vast inspection of the traffic accident scene, accident analysis, real vehicle road tests for drivers' emergency operation behavior, tests for drivers' behavior and psychological assessment, theoretical analysis, and by building the algorithm-analysis model of drivers' operation ability to retreat from risk and vehicle response extent, data-image real-time recognition-calculation model for vehicle running environment (road, signs, marked line, front vehicle conditions) and algorithm-analysis model for driving path image recognition, running condition detection and motion track prediction, define mode and identification mode of drivers' dangerous driving behavior is studied with the vehicles operated and drove. Through this project, it is possible to improve road traffic transportation safety, manage and normalize drivers' operation behavior with efficiency, and improve the technological content and capacity of road traffic

英文关键词： driver；dangerous driving；driving behavior；machine vision；identification model