| 高分辨率遥感植被分类研究 |
| |
| 引用本文： | 陈君颖,田庆久.高分辨率遥感植被分类研究[J].遥感学报,2007,11(2):221-227. |
| |
| 作者姓名： | 陈君颖 田庆久 |
| |
| 作者单位： | 南京大学,国际地球系统科学研究所,江苏,南京,210093;中国科学院,中国遥感卫星地面站,北京,100086 |
| |
| 基金项目： | 江苏省高技术研究发展计划项目 |
| |
| 摘 要： | 以南京市区的植被覆盖为研究对象,基于IKONOS遥感影像,采用决策树分类算法,根据各种植被光谱特征建立知识库,提出基于光谱信息的植被分类方法,继而结合高分辨率影像特有的纹理特征引进局部一致性指数对该方法进行改进,提出结合纹理信息的高分辨率遥感植被分类方法,分类总体精度从仅利用光谱信息的83.16%显著提高到91.89%,Kappa系数达到0.8886。采用Quickbird遥感影像对该方法进行验证,分类总体精度为91.94%,Kappa系数为0.8783,表明该植被分类方法能有效地对植被进行分类与识别,精度较高,且对于不同数据源的植被分类具有一定的普适性,为实现植被的自动化提取提供了理论依据和有效的方法途径。
|
| 关 键 词： | 高分辨率遥感 植被分类 知识库 决策树 纹理 |
| 文章编号： | 1007-4619（2007）02-0221-07 |
| 修稿时间： | 2005-08-282006-04-06 |
Vegetation Classification Based on High-resolution Satellite Image |
| |
| CHEN Jun-ying and TIAN Qing-jiu.Vegetation Classification Based on High-resolution Satellite Image[J].Journal of Remote Sensing,2007,11(2):221-227. |
| |
| Authors: | CHEN Jun-ying and TIAN Qing-jiu |
| |
| Institution: | 1.International Institute for Earth System Science, Nanjing University, Jiangsu Nanjing 210093, China; 2. China Remote Sensing Satellite Ground Station, Beijing 100086, China |
| |
| Abstract: | This paper used Nanjing city as the research object for vegetation classification based on IKONOS image. We adopted classification algorithm of decision tree, built knowledge pool on the basis of vegetation spectral information and presented a vegetation classification method based on the spectral information. Then the method was improved by combining textural information and using Homogeneity index. The overall precision of the vegetation classification method based on spectral information was 83.16% , but that of the improved method was upgraded to 91.89% , and kappa coefficient 0. 8886. A Quickbird image was used to validate the method, the overall precision was 91.94% and kappa coefficient 0. 8783. The result showed the vegetation classification method can classify and discriminate vegetation effectively, and the precision is high. As to vegetation classification with different data source, the method has universality. The vegetation classification method provides theoretical foundation and effective method for extracting vegetation automatically. |
| |
| Keywords: | high-resolution remote sensing vegetation classification knowledge pool decision tree texture |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
| 点击此处可从《遥感学报》浏览原始摘要信息 |
| 点击此处可从《遥感学报》下载免费的PDF全文 |
|
