基于机器视觉的车道线识别算法  

Algorithm of lane recognition based on machine vision

作　　者：牛明花 张建民[1] 李根[1] 

机构地区：[1]天津职业技术师范大学电子工程学院,天津300222

出　　处：《天津职业技术师范大学学报》2015年第3期16-19,共4页Journal of Tianjin University of Technology and Education

摘　　要：针对传统车道线检测识别算法运算复杂、处理速度较慢以及鲁棒性欠缺等问题,提出了一种新的快速车道线识别算法。将摄像机采集到的道路图像进行预处理,对二值化的3种算法进行图像仿真实验,对传统的边缘检测Sobel算法进行改进,利用霍夫变换(Hough Transform)进行车道线识别。仿真实验表明:该算法达到了快捷准确的识别效果。In order to overcome the shortcomings of previous lane detection algorithms in computational complexity, speed and robustness, a new and fast lane detection algorithm is proposed in this paper. The road image captured by the camera is preprocessed, image simulation experiments of 3 kinds of binarization algorithms are made, the traditional Sobel algorithm in edge detection is developed and the Hough Transform algorithm is used to detect the lane. The experiments show that this al- gorithm achieves the expected effect.

关 键 词：车道线识别 最大类间方差法 SOBEL算子 霍夫变换 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] U463.6[自动化与计算机技术—计算机科学与技术]