一种改进Hough变换的车道线检测算法  

An Improved Algorithm for Lane Detection Based on Hough Transformation

作　　者：王全[1] 

机构地区：[1]西安工业大学计算机科学与工程学院,西安710032

出　　处：《计算机与数字工程》2014年第11期2164-2167,共4页Computer & Digital Engineering

基　　金：西安工业大学校长基金项目(编号:605-01001316)资助

摘　　要：对基于单目视觉的车道线检测技术进行研究,针对复杂环境中车道线检测易受图像中其它直线段的影响而出现较多误检测的问题,提出一种改进算法。首先将车道线图像灰度化,采用中值滤波去除图像噪声;再次,采用Canny算子和OTSU方法进行边缘增强和图像二值化;最后,使用改进的Hough变换进行车道线检测。实验结果表明,该方法可以有效地滤除其它直线段的影响,在复杂环境中获得良好的车道线检测效果。Lane detection technology based on the visual image is researched. The lane detection in complex environment is easily affected by the inmge of the other straight lines, thus giving rise to errors. In this paper, an improved algorithm for lane detec- tion is put forward. Firstly, the color image is turned into grayscale and noise is filtered out by median filter. Secondly The Canny method is used for edge enhancement and the OTSU algorithm is used to do the binarization on lanes images. Finally the improved Hough transfom~ is used to detection the lane. The experimental results show that this method can effectively remove the interference caused by other lines and has the excellent lane line detection results under complex environment.

关 键 词：车道线检测 HOUGH变换 CANNY算子 OTSU阈值 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] U461.6[自动化与计算机技术—计算机科学与技术]