摘要
为了提高智能车辆道路识别的准确性和效率,设计了一种基于平行边缘特征的道路检测算法。提出了基于边缘连接的道路区域快速粗分割方法;对边缘点局部直线的检测和方向进行编码,利用竖直线实现了极大可能道路区域的估计;基于方向一致性判别准则,实现了极大可能道路区域内平行边缘的识别算法;提出了三个道路识别准则,综合运用平行边缘、道路的区域位置信息,实现了道路特征的准确识别。实验结果表明,本文算法能够快速并准确的提取典型的直线和弯曲道路模型中的道路区域,比以往算法在速度和准确性上都有较大的提升。
In order to improve the accuracy and efficiency of intelligent vehicle road extraction, a road detection algorithm based on parallel edge features is proposed. A new fast rough segmentation algorithm for road area image based on edge connection is proposed. Edges points local linear detection and direction are coded, and the vertical lines are detected to estimate the most likely road region. The parallel edges in the most likely road region are extracted based on the similarity degree of local orientation, and three efficient criterions are presented to extract the road region. Experimental results show that the proposed algorithm is suitable for various road images, which can accurately detect the road regions of straight and curved roads at a fast time. The speed and accuracy of the proposed algorithm improves largely.
出处
《光学学报》
EI
CAS
CSCD
北大核心
2015年第7期224-230,共7页
Acta Optica Sinica
基金
国家自然科学基金(61005034
51175449)
河北省自然科学基金(F2012203185)
关键词
机器视觉
平行边缘识别
方向编码
特征识别
machine vision
parallel edges detection
road detection
feature recognition
