结构化道路中的车道偏离预警系统研究

发布时间：2018-07-03 17:10

本文选题：计算机视觉 + 车道偏离预警　；参考：《山东大学》2014年硕士论文

【摘要】：随着道路交通的发展和人们汽车拥有量的增加,交通事故的发生率也越来越高,行车安全问题越来越受到人们的重视。针对这一问题,结合当前科技的高速发展,人们开始寻求一种能够利用计算机科技来减少交通事故的安全辅助驾驶系统(DAS),而基于计算机视觉的车道偏离预警系统则是其中最为重要的一部分。车道偏离预警系统旨在及时提醒驾驶员因疲劳驾驶而导致的车辆偏离正确行驶路线,以便驾驶员及时做出纠正,保持车辆在正确的车道内行驶。车道偏离预警系统的基础是车道线的识别与跟踪,即首先要通过摄像机获得行驶道路的车道线信息,以便系统做出判断,是否偏离当前车道,决定是否给驾驶员提醒。本文首先介绍了车道线与跟踪中的基本步骤,研究了如何快速有效的获取道路图像中的车道线信息,提出了自己的想法并进行了实验验证,最后对整个车道偏离预警系统进行了实现。文章主要分为以下几个部分： (1)介绍了基于计算机视觉的车道偏离预警系统这一课题的研究背景和意义,介绍了国内外发展现状,并对整个课题进行了系统的分析。 (2)介绍并分析了图像预处理各个模块中所用到的相关方法,并根据道路图像处理的具体要求,选择了合适的算法并有针对的进行了改进(采用中值滤波进行图像增强,改进的Sobel算子进行边缘检测,图像二值化选用的是大津法)。 (3)首先选择了车道线识别的模型,并通过对道路图像的分析划分了车道线识别与跟踪的感兴趣区域。在车道线初始检测阶段和车道线跟踪阶段选取了不同的车道线识别方法(改进的Hough变换与最小二乘法相结合),即保证了车道线识别的鲁棒性又提高了车道线识别的实时性。 (4)介绍了几种车道偏离预警的模型,选取了合适的模型(TLC模型)并提出了具体有效的车道线偏离的判定标准。 (5)总结与展望。对本文的研究工作进行了总结,分析了所用方法的优点和不足,并对下一步工作进行了展望。
[Abstract]:With the development of road traffic and the increase of people's car ownership, the incidence of traffic accidents is becoming higher and higher. People pay more and more attention to the problem of traffic safety. In view of this problem, combined with the rapid development of current science and technology, People begin to seek a kind of safety assistant driving system (DAS) which can use computer technology to reduce traffic accidents, and the driveway deviation warning system based on computer vision is the most important part. The purpose of lane deviation warning system is to remind drivers that the vehicles caused by fatigue drive deviate from the correct route in time so that the drivers can make corrections in time and keep the vehicles in the right driveway. The roadway deviation warning system is based on lane line recognition and tracking, that is, the first step is to obtain lane information through the camera, so that the system can judge whether the lane deviates from the current lane and decide whether to remind the driver. This paper first introduces the basic steps of lane line and tracking, studies how to get lane information from road image quickly and effectively, and puts forward some ideas and experiments. Finally, the whole lane deviation warning system is realized. This paper is mainly divided into the following parts: (1) introduce the research background and significance of the driveway deviation early warning system based on computer vision, introduce the domestic and foreign development status, And the whole subject is systematically analyzed. (2) the relevant methods used in each module of image preprocessing are introduced and analyzed, and according to the specific requirements of road image processing, The appropriate algorithm is selected and improved (median filter is used for image enhancement, improved Sobel operator is used for edge detection, and Otsu method). (3 is used for image binarization). First, the model of lane line recognition is selected. Through the analysis of the road image, the area of interest in lane line recognition and tracking is divided. Different lane line recognition methods (improved Hough transform and least square method) are selected in the initial detection stage of lane line and lane track tracking stage, which ensures the robustness of lane line recognition and improves lane line recognition. (4) several models of lane deviation warning are introduced. The suitable model (TLC model) is selected and the criterion of lane deviation is put forward. (5) Summary and prospect. The research work of this paper is summarized, the advantages and disadvantages of the methods are analyzed, and the future work is prospected.
【学位授予单位】：山东大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP391.41;U495

【相似文献】