학술논문

基于Hough变换的城市环境道路识别优化算法研究 / Research on improved urban environment road detection algorithm based on Hough transform
Document Type
Academic Journal
Source
机械设计与制造工程 / MACHINE DESIGN AND MANUFACTURING ENGINEERING. 46(12):71-75
Subject
单目视觉
车道线检测
感兴趣区域
改进Hough变换
monocular vision
lane detection
region of interest ( ROI)
progressive Hough transformation
Language
Chinese
ISSN
2095-509X
Abstract
在单目视觉的前提下,针对复杂多变的道路环境中车道线识别算法易受干扰的现状,设计了一种新的道路图像处理方法.首先静态划分道路图像的感兴趣区域,在感兴趣区域内进行图像处理和车道线检测,将图像灰度化,采用中值滤波去除噪声;接着在基于最大类间方差法的基础上,运用Canny算子进行图像边缘检测;最后使用改进的Hough变换对车道线进行识别及拟合.通过自主搭建的实验平台对研究算法进行实验检测,实验结果表明,该方法在道路环境复杂的情况下仍能准确地识别车道线,具有较强的鲁棒性.
The lane detection algorithm is easy to be interfered by the complex road environment .Aiming at this problem it proposes a reasonable road detection method .It firstly extracts the static region of interest ( ROI ) , and realizes the image processing and lane detection in this RIO .Then it makes the road image to the grey scale image, and eliminates the noise in median filter .Using the Canny edge detector to detect the edge of image based on OTSU , it realizes the lane detection in the progressive Hough transformation .The experiment result shows that this method can accurately identify the lane line even in the complex road environment , and have good robustness .