摘要
结合图像中直线段点密度分布的一些视觉经验,提出一种检测线段的点对分析算法,避免了Hough变换中累加参数空间的复杂计算过程。实验表明,本文方法具有精度高、消耗内存量小并能直接检测出线段及其端点的优点,可应用在弹道轨迹检测、字符笔画识别和遥感图像道路检测等领域。
This paper introduce a new way of detection lines by analyzing point pair in the image. Much time on amounting the parameter space is saved. Furthermore, the human experiments are also consult in this new method. The end points of each line can also be detected accurately with low cost of memory and time. The paper gives its application in trajectory detection, character recognition and road detection.
出处
《武汉大学学报(信息科学版)》
EI
CSCD
北大核心
2008年第3期314-317,共4页
Geomatics and Information Science of Wuhan University
基金
国家自然科学基金资助项目(60705009)
关键词
线段检测
HOUGH变换
弹道检测
字符识别
line detection
Hough transform
trajectory detection
character recognition