首页|期刊简介|投稿指南|分类索引|刊文选读|订阅指南|资料|样刊邮寄查询|常见问题解答|联系我们
王智宇,陈光,陈勇. 基于OpenCV的改进RANSAC车道线检测方法[J]. 科学技术与工程, 2019, 19(28): 222-226.
wangzhiyu,chenyong.Improved RANSAC Lane Detection Method Based on OpenCV[J].Science Technology and Engineering,2019,19(28):222-226.
基于OpenCV的改进RANSAC车道线检测方法
Improved RANSAC Lane Detection Method Based on OpenCV
投稿时间:2019-03-22  修订日期:2019-05-05
DOI:
中文关键词:  车道线检测 积分图 Harris角点 RANSAC
英文关键词:lane detection integral image harris corner ransac
基金项目:
        
作者单位
王智宇 河北工业大学机械工程学院
陈光 河北工业大学机械工程学院
陈勇 河北工业大学机械工程学院
摘要点击次数: 232
全文下载次数: 51
中文摘要:
      车道线的检测技术是自动驾驶汽车中的重要技术。为了提高车道线的检测能力,本文提出了一种改进RANSAC的车道线识别方法。该方法通过设置感兴趣区域提取路面图像并进行缩放;把彩色图像的RGB通道按5:5:0的权重转化成灰度图像;再用速度更快的积分图法对图像进行自适应二值化;接下来进行一系列的形态学处理来减小噪声;提取Harris角点作为拟合数据点;最后,运用改进了选择初始点和删除外点的RANSAC的方法,根据数据点估计车道线参数。实验结果表明,该算法适合多种道路环境下的车道线检测,具有较好的鲁棒性和实时性。
英文摘要:
      Lane detection technology is an important technology in Autonomous vehicles. In order to improve the ability of lane detection, an improved RANSAC lane detection method was proposed in this paper. In this method, the region of interest which involves lane of an image was set. Then the region of interest was scaled. The color image was converted into a gray-scale image by weighted-summing RGB values with 5:5:0 weight. Then, a faster adaptive integral image method was used to binarize the image. Next, a series of morphological treatments were taken to reduce noise. Harris corner points were extracted as fitting data points. Finally, the improved RANSAC method which selects the initial points and deletes the outer points was employed to estimate the lane parameters based on the data points. The experimental results show that the proposed algorithm is suitable for lane detection in a variety of road environments and has good robustness and real-time performance.
查看全文  查看/发表评论  下载PDF阅读器
关闭
你是第31873988位访问者
版权所有:科学技术与工程编辑部
主管:中国科学技术协会    主办:中国技术经济学会
Tel:(010)62118920 E-mail:stae@vip.163.com
京ICP备05035734号-4
技术支持:本系统由北京勤云科技发展有限公司设计

京公网安备 11010802029091号