测绘科学技术学报Issue(1)：53-56,4.DOI:10.3969/j.issn.1673-6338.2014.01.012
结合形状与同质特征的遥感影像城市道路提取方法
A Method of Urban Road Extraction from Remote Sensing Image Combining Shape Features and Homogeneity
摘要
Abstract
Road extraction from remote sensing imagery based on spectral or geometric information only is difficult to separate urban road features from background completely because there are a lot of noises such as shadow building vehicle etc in the remote sensing imagery. To overcome the aforementioned shortcomings of the common existing road extraction algorithms a new method to extract the urban road from remote sensing imagery is presented in the paper. Firstly remote sensing imagery is homogenized by using gradient image. Then noises are e-liminated by using shape features to highlight the road information. As a final step these seed segments can be re-constructed to generate a proper road map based on relational constraints of perceptual organization. The approach which takes QucikBird imagery as sample data is shown to perform well with urban road extraction.关键词
形状特征/同质特征/知觉组织/遥感影像/道路提取Key words
shape features/homogeneity/perceptual organization/remote sensing image/road extraction分类
天文与地球科学引用本文复制引用
许锐..结合形状与同质特征的遥感影像城市道路提取方法[J].测绘科学技术学报,2014,(1):53-56,4.基金项目
福建省教育厅科技项目( JA13220)。 （ JA13220）
