摘要
针对卫星图像中道路的基本特征以及道路信息提取的难点,提出了一种基于区域生长型分水岭算法的卫星图像主干道路提取方法。该方法首先运用高通滤波器对原始图像进行去噪和平滑,然后采用区域生长型分水岭算法提取道路信息,最后用目标面积特征将误提的小斑点去除。实验结果表明,该方法性能良好,可以取得很好的效果。
In view of the basic characteristic of road and the difficulty in road information extraction, a region-growth watershed algorithm for road extraction from satellite image is proposed. Firstly high-pass filter method is used to de-noise and smooth original image, and then region-growth watershed algorithm is used to extract road information, and finally some non-road object is removed by area character. The experiment shows this method is efficient.
出处
《计算机工程与设计》
CSCD
北大核心
2008年第19期4987-4988,5013,共3页
Computer Engineering and Design
关键词
卫星图像
道路提取
高通滤波器
分水岭
区域生长
satellite image
road extraction
high-pass filter
watershed
region growth