中文摘要：

以我国自主发射的第一颗高分辨率卫星（GF-1卫星）的影像为数据源，在东洞庭湖开展湿地信息提取。结合国家湿地遥感分类系统和高分影像数据的可解译性，建立了东洞庭湖湿地类型遥感分类系统，包括湖泊、河流、泥滩地、林地、芦苇、苔草、辣蓼＋泥蒿7种湿地类型。利用最佳指数和联合熵2个指标确定最佳波段组合方式为4（R）3（G）2（B），并在遥感图像进行预处理的基础上，采用面向对象的决策树进行分类，结果表明：基于面向对象决策树遥感分类方法比最大似然法分类的精度高，总分类精度达85.41%，Kappa系数为0.8225，并且各类信息的提取精度均有所提高，其中湖泊、芦苇、林地的精度较高。

英文摘要：

Based on the first high satellites （GF-1 satellite images） launched by China, the study extracted information of the East Dongting Lake wetland. Combined with the interpretation of national wetland remote sensing classiifcation system and high resolution image datas, a RS classiifcation system of wetland types of East Dongting Lake was built , including lakes, rivers, mudflats, woodland, reed, sedge, Polygonum ＋ mud Artemisia seven kinds of wetlands .Two index ,the best index and joint entropy, were used to determine the optimal band combination ,that was 4 （R） 3 （G） 2 （B）.Based on the preprocessing of remote sensing image,the classiifcation was done by using object oriented decision tree, the results showed that： the precision of object oriented decision tree classiifcation method is higher than maximum likelihood ratio classiifcation, the accuracy of total classiifcation reaches to 85.41%, Kappa coefifcient is 0.8225. And the extraction accuracy of all kinds of information has improved, among them the accuracy of lake, reed and forest is higher.