摘要
研究基于Landsat7 Enhanced Thematic Mapper Plus(ETM+)影像,利用典型地物的光谱特征及灰度共生矩阵的纹理特征建立分类模型,提取了微山湖地区的泥滩信息,并采用随机采样法对分类结果进行了评价,Kappa系数达到0.822 7.研究表明,单一使用光谱特征分类的方法不能够有效的区分水体与泥滩信息,而辅以纹理特征后的分类方法可以将二者区分开,并且能够达到一定的分类精度,为湿地遥感信息的分类提供了一种方法.
Based on Landsat7 Enhanced Thematic Mapper Plus （ ETM ＋ ） image data, the information of mudflats of Weishanhu Lake area was extracted, by using of established classification model method with spectral characteristics and gray level concurrence matrix texture feature of typical types. In addition, random sampling method was evaluated on the classifications with an overall accuracy （ Kappa Coefficient） of 0. 8227. The results show that spectral features classification method which is single used can＇t effectively separate the information of mudflats from wetland. While it is combined with the texture feature classification method, it could separate them well and achieve a certain classification accuracy . So it provides a method for the extraction of remote sensing information.
出处
《湖南科技大学学报（自然科学版）》
CAS
北大核心
2011年第4期68-72,共5页
Journal of Hunan University of Science And Technology：Natural Science Edition
基金
湖南省自然科学基金项目(10JJ4027)
湖南省教育厅一般项目(10C0722)
