【摘要】 城镇地物类型复杂多样,道路狭窄且易与周围地物混合,这导致城镇背景下的道路信息提取存在困难。该文以安徽省滁州市部分区域为实验区,使用资源三号卫星影像,采用面向对象的方法,利用光谱特征、形状特征等信息,在对象级基础上通过复合向量机的方法进行城镇背景下的道路信息循环拾取,并利用形态学方法对提取的结果进行了分类后处理。研究结果表明,该文提出的道路提取方法具有可行性。更多还原

【Abstract】 Roads is characterized by rapid update in an urban context.This makes it significant that the information on road is extracted quickly and precisely.The roads in urban areas are characterized as being complex,various and sometimes narrow.This makes it difficult to extract road features in an urban context due to the existence of a large number of mixture pixels.In this paper,apart of Chuzhou,Anhui is taken as test area and Ziyuan 3imagery is used to explore the effective method of road extraction in an urban context.Based on object-oriented image analysis,multiple features are fed into support vector machine for the extraction of roads in an urban context.These features include spectral features,shape feature and other features.Finally,the results are processed by using morphology method.The research shows that the proposed method is effective and road extraction results are satisfying.更多还原