文献类型：期刊文献

中文题名：基于“高分四号”卫星影像洞庭湖湿地信息提取

英文题名：Comparative Study of Dongting Lake Wetland Information Extraction Based on GF-4 Satellite Image

作者：由佳[1] 张怀清[1] 陈永富[1] 刘华[1] 高志海[1]

第一作者：由佳

机构：[1]中国林业科学研究院资源信息研究所

年份：2016

卷号：37

期号：4

起止页码：116-122

中文期刊名：航天返回与遥感

外文期刊名：Spacecraft Recovery & Remote Sensing

收录：CSTPCD;;北大核心:【北大核心2014】；

基金：基金项目:国家重大专项(21-Y30B05-9001-13/15-2)

语种：中文

中文关键词：湿地;信息提取;洞庭湖;“高分四号”卫星;卫星应用

外文关键词：information Extraction; Dongting Lake; GF-4 satellite; satellite application

分类号：TP75

摘要：文章以"高分四号"卫星数据为数据源,以湖南洞庭湖区域为主要研究区域,建立了洞庭湖三级湿地分类系统(主要包括:水体、草滩地、泥滩、耕地、林地、裸地六类),分别使用最大似然法、支持向量机法以及神经网络法对洞庭湖湿地进行遥感信息提取。结果显示,在三种不同的遥感分类方法中,最大似然法分类精度相对较高,总体精度为77.14%,Kappa系数为0.592 9。研究发现:"高分四号"卫星影像具有良好的湿地类型信息提取能力,采用最大似然分类方法可以较好的进行大区域湿地类型信息的提取。
With GF-4 Satellite data as the data source and Dongting Lake area in Hunan Province as the main research area, this paper establishes a classification system for Dongting Lake level-3 wetlands, mainly including six types, namely, water body, grass land, mud flat,cultivated land, forest land and bare land, and uses the methods of maximum likelihood, support vector machine and neural network for the extraction of remote sensing information of Dongting Lake wetlands. The results show that the maximum likelihood method has a relatively high classification accuracy with an overall accuracy of 77.14% and a Kappa coefficient of 0.592 9 among the above three different remote sensing classification methods. According to the research findings, the ＂GF-4 Satellite＂ image has a large capability of wetland type information extraction, and the use of maximum likelihood classification method can lead to better extraction of large-area wetland type information.