摘要
针对作者自主开发的视觉导航区域交通智能车辆(Cyber Car)的导航路径在阴影条件下难以应用单一传统阈值方法分割提取的问题,提出了基于区域生长的图像分割方法。该方法利用路径边缘信息寻找种子点,进而对整个路径区域进行生长,最终实现车辆导航路径的准确分割与提取。同时为满足车辆导航实时性的要求,还提出了对分块子图像进行压缩的方法。对不同阴影下导航路径图像的分割试验表明,该方法能够实时、准确地识别导航路径,并具有较强的抗干扰能力。
In order to solve the problem that the traditional single threshold method can not segment the shadow path mark image successfully, a new algorithm based on Region-growing has been proposed for our indigenously designed and manufactured vision navigation intelligent vehicle （Cyber Car）. It utilizes the path mark edge to search several kernel points. Therefrom the path mark region grows and the segmentation and extraction of the navigation path mark can be achieved. The divided sub-images compression method was proposed to fulfill the real-time demands. The segmentation experiment results of different shadow path mark images show that the new method can segment the navigation path accurately and quickly with good antijamming ability.
出处
《吉林大学学报（工学版）》
EI
CAS
CSCD
北大核心
2006年第B03期132-135,共4页
Journal of Jilin University:Engineering and Technology Edition
基金
中国博士后科学基金资助项目(2004036397)吉林大学青年教师基金资助项目
关键词
视觉导航
图像分割
区域生长
阴影
vision navigation
image segmentation
region-growing
shadow
