基于Landsat OLI数据的不透水面提取方法对比研究——以长株潭城市群核心区为例

Compared of Impervious Surface Extraction Methods based on Landsat8 OL——A Case Study of the Core Area of Changsha-Zhuzhou-Xiangtan Urban Agglomeration

不透水面的分布规模和空间格局是影响区域生态系统和环境变化的重要指标。该研究基于LandsatOLI遥感影像,以长株潭主城区核心区为研究区域,进行了归一化建筑指数法、归一化差值不透水面指数法和最小噪声分离结合支持向量机分类法3种方法提取不透水面信息的综合比较。结果表明:MNF结合SVM方法步骤相对较多,在选取端元和去除水体掩膜等方面都存在人工干涉,受人工影响较大;归一化建筑指数法虽然能快速地提取不透水面,但存在裸土与不透水面混淆的情况;而归一化差值不透水面指数法可有效区分裸土与不透水面,提取精度高于其他2种方法。最终选择归一化差值不透水面指数法,提取研究区2013年、2015年的不透水面信息,分析了其空间格局和扩展情况。

lmpervious surface of distribution scale,spatial framework and diversification are the key indicators of regional ecosystem and environmental change.Remote sensing technology based on OLI image,the information of impervious surface is extracted,and the normalized building index method is compared and analyzed by taking the core area of Changsha-Zhuzhou-Xiangtan City as the research area.The combination of normalized difference impervious surface index method and minimum noise separation support vector machine classification.The results showed that,the combination of MNF and SVM method steps are relatively more complex.In the selection of end member and removal of water Mask,artificial interference exists,which is greatly by artificial.The normalized building index method can quickly extract impervious surface,but the situation of bare land and construction land is confused.The normalized difference impervious surface index method can solve this problem.And the precision of extracting the construction land is higher than that of the normalized building index method.