文献类型：期刊文献

中文题名：基于Worldview-2八波段影像改进指数的湿地类型分类研究

英文题名：Research on Wetland Type Classification Based on Improved Remote Sensing Index of Worldview-2 Data

作者：凌成星[1] 鞠洪波[1] 张怀清[1] 孙华[1]

第一作者：凌成星

机构：[1]中国林业科学研究院资源信息研究所

年份：2014

卷号：27

期号：5

起止页码：639-643

中文期刊名：林业科学研究

外文期刊名：Forest Research

收录：CSTPCD;;Scopus;北大核心:【北大核心2011】；CSCD:【CSCD2013_2014】；

基金：国家高技术研究发展计划(863计划)(2012AA102001);国家重大专项(E0305/1112/02)

语种：中文

中文关键词：湿地;Worldview-2影像;遥感信息提取;改进遥感指数

外文关键词：wetland; Worldview-2 image; remote sensing information extraction; improved remote sensing index;

分类号：S725.7

摘要：采用Worldview-2八波段影像作为数据源,选取东洞庭湖湿地核心区域作为研究区,进行了Worldview-2八波段特征分析、构建改进遥感指数、采用改进遥感指数阈值分层分类的策略对湿地区域进行信息提取。研究结果表明,基于Worldview-2八波段影像改进指数的湿地类型分类总精度达到了92.24%,Kappa系数为0.902,比原始遥感指数的分类精度提高了8.18%,特别是对草滩地和泥滩地的区分有了较大的提高,是有效、准确提取湿地类型的技术方法。
Taking the core area in wetland nature reserve of Dongting Lake in Hunan province as the reserch plot, the high resolution image Worldview-2 data were analyzed and used to establish the improved remote sensing index. The information about the wetland was extracted with the strategy of data threshold hierarchical classification. The result showed that accuracy of wetland type classification reached 92.24%, with Kappa coefficient at 0.902. The accuracy of classification, especially for processing the data of marsh , meadow and bare land, was 8.18% higher compared to the original remote sensing index method. It is proved that the improved remote sensing index extration mothod is an effective way in wetland remote sensing data classification.