基于图像处理的汽车车道偏离预警系统研究与实现

发布时间：2019-01-10 12:10

【摘要】：随着当今社会经济迅速发展,人们生活水平逐渐提高,汽车的数量也飞速增加。而汽车数量的大量增加,使得交通环境日益恶劣,不仅使交通拥堵,还带来越来越多的交通事故。为了减少交通事故的发生,相关研究人员提出了智能交通系统的思想,而车道偏离预警系统是智能交通系统的核心。车道偏离预警系统是一种通过对摄像头采集的道路图像进行检测得到当前车辆位置的信息,如果车辆处于车道偏离的状态则发出预警信号提醒驾驶员安全驾驶的系统。本文通过对现有的国内外车道偏离预警系统进行简单的研究分析,确定了本文的研究内容,主要分为三个模块即图像预处理模块、车道线检测与跟踪模块以及车道偏离预警模块,本论文的主要研究内容和贡献如下:1、对摄像头采集到的原始道路图像进行相应的预处理步骤,主要包括道路图像灰度化、对道路图像进行滤波增强以及道路图像的边缘检测,对比分析几种不同的图像预处理算法,结合道路图像的实际需求选取合适的处理方法,并选取直线模型作为本文的车道线模型,通过分析几种经典的直线检测方法并比较各自的优缺点,选取Hough变换来对道路图像中的车道线进行直线检测。2、在车道线的检测过程中,提出一种Hough变换结合灰度块匹配的车道线检测方法,图像经Hough变换后检测出多条直线,然后利用道路灰度和车道线灰度的明显差异,对其进行灰度块匹配从而在多条直线中确定真正的左右车道线,并利用Kalman滤波对车道线进行实时跟踪。3、对几种经典的车道偏离预警模型简要分析,结合本文的实际情况提出一种基于车道偏离率结合直线数量判断的车道偏离模型,首先根据左右车道线的角度计算车道偏离率,在此基础上增加对检测直线的数量进行判断的条件,从而对车辆当前状态进行准确判断,以达到系统准确预警的功能需求,并对不同场景下的车辆偏离情况进行具体实验分析。通过大量实验及实验结果分析,本文所用的车道线检测方法和车道偏离预警方法效果较好,研究实现的车道偏离预警系统是有效的。
[Abstract]:With the rapid development of social economy, people's living standard is improving gradually, and the number of cars is also increasing rapidly. With the increase of the number of cars, the traffic environment is getting worse and worse, which not only makes the traffic jam, but also brings more traffic accidents. In order to reduce the occurrence of traffic accidents, some researchers put forward the idea of Intelligent Transportation system (its), and the driveway deviation early warning system is the core of its. The lane deviation warning system is a system that detects the current vehicle position by detecting the road image collected by the camera. If the vehicle is in the driveway deviation state, it sends out an early warning signal to remind the driver to drive safely. Through the simple research and analysis of the existing driveway deviation warning system at home and abroad, this paper determines the research content of this paper, which is mainly divided into three modules, that is, image preprocessing module. The main research contents and contributions of this paper are as follows: 1. The corresponding preprocessing steps of the original road image captured by the camera include grayscale road image. The road image is enhanced by filtering and edge detection of road image. Several different image preprocessing algorithms are compared and analyzed. According to the actual needs of road image, the appropriate processing method is selected. The line model is selected as the lane line model in this paper. By analyzing several classical line detection methods and comparing their advantages and disadvantages, Hough transform is selected to detect the lane line in the road image. 2. In the course of lane line detection, a method of lane line detection based on Hough transform and grayscale block matching is proposed. After Hough transform, many straight lines are detected, and then the obvious difference between road gray and lane gray is used. The grayscale block matching is carried out to determine the true left and right lane lines in many straight lines, and the lane line is tracked in real time by using Kalman filter. 3. Several classical lane deviation warning models are briefly analyzed. According to the actual situation of this paper, a lane deviation model based on lane deviation rate combined with the number of straight lines is proposed. Firstly, the lane deviation rate is calculated according to the angle of the left and right lane lines. On this basis, the condition of judging the number of detection lines is added, so that the current state of the vehicle can be accurately judged, so as to meet the functional requirements of accurate warning system, and the deviation of vehicles under different scenes is analyzed through specific experiments. Through a large number of experiments and analysis of experimental results, the lane line detection method and lane deviation warning method used in this paper have a good effect, and the developed lane deviation warning system is effective.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41;TP277

【相似文献】