| 基于多特征的遥感影像决策树分类 |
| |
| 引用本文： | 潘琛.基于多特征的遥感影像决策树分类[J].光电子．激光,2010(5):731-736. |
| |
| 作者姓名： | 潘琛 |
| |
| 作者单位： | 同济大学测量与国土信息工程系遥感与空间信息技术研究中心; |
| |
| 基金项目： | 上海市科学技术委员会科研计划资助项目(08DZ1205708); 世界自然基金会科研资助项目(CN08791-2.4.02.04) |
| |
| 摘 要： | 构建了一种基于多特征的遥感影像决策树分类方法。通过对遥感影像进行波段代数运算、主成分分析和图像分割等处理,提取出影像上地物的光谱维特征、纹理特征和形状特征。在此基础上,结合试验区主要地物类型提纯后的训练样本集,采用C5.0决策树分类法进行影像分类,实现主要地物的空间分布专题信息提取,并利用该方法对Landsat-5TM影像进行了分类实验。结果表明,所提出的方法能够有效地提高分类精度。
|
| 关 键 词： | 多特征 决策树分类 遥感影像 形状特征 |
Decision tree classification of remote sensing images based on multi-feature |
| |
| PAN Chen.Decision tree classification of remote sensing images based on multi-feature[J].Journal of Optoelectronics·laser,2010(5):731-736. |
| |
| Authors: | PAN Chen |
| |
| Affiliation: | PAN Chen,LIN Yi,CHEN Ying-ying(Research Center of Remote Sensing , Spatial Information Technology,Department of Surveying , Geoinformatics,Tongji University,Shanghai 200092,China) |
| |
| Abstract: | A classification method is presented for remote sensing images,which uses C5.0 decision tree classifier based on multi-feature.The features of spectrum,texture and shape in the image are extracted with a series of operations such as algebraic manipulation of wave bands,principal component analysis,image segmentation,and so on.Then,the extraction of the thematic information of primary ground-object distribution is realized through a combination of the purified training samples for each class and the image cl... |
| |
| Keywords: | multi-feature decision tree classification remote sensing images shape feature |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《光电子．激光》浏览原始摘要信息 |
|
点击此处可从《光电子．激光》下载全文 |
|
