摘要
研究了基于结构张量的主方向计算方法,针对其计算结果不够精确的缺点,对高斯滤波进行改进并结合canny算子,提出了改进的主方向计算方法,然后基于局部主方向结合Gibbs抽样进行遥感图像道路提取。截取某城市Quick Bird高分辨率遥感图像中的两幅图片,对本算法进行测试。结果表明,该算法可以比较精确地对道路进行提取。
This paper studies the main direction calculation method based on the structure tensor,The calculation result of which is not accurate enough.Gauss filter is improved and combined with Canny operator,an improved main direction algorithm is proposed.Then based on the local main direction with Gibbs sampling,road is extracted from remote sensing images.With Two images intercepted from a Quick Bird high resolution remote sensing images of a city,the algorithm is tested.the results show that the algorithm can extract road accurately.
出处
《科技通报》
北大核心
2013年第2期175-177,共3页
Bulletin of Science and Technology