基于S~3VM模型的高光谱遥感影像分类 |
| |
引用本文: | 魏立飞,俸秀强,李丹丹,牟紫薇.基于S~3VM模型的高光谱遥感影像分类[J].测绘通报,2017(12):43-47. |
| |
作者姓名: | 魏立飞 俸秀强 李丹丹 牟紫薇 |
| |
作者单位: | 1. 湖北大学资源环境学院,湖北武汉430062;区域开发与环境响应湖北省重点实验室,湖北武汉430062;2. 湖北大学资源环境学院,湖北武汉,430062;3. 农业部农业信息技术重点实验室,北京,100081 |
| |
基金项目: | 国家自然科学基金,干旱气象科学研究基金,数字制图与国土信息应用工程国家测绘地理信息局重点实验室开放研究基金 |
| |
摘 要: | 针对传统的高光谱遥感影像分类受限于训练样本的个数,难以取得较好分类结果的不足,提出了一种基于聚类核的半监督支持向量机(S3VM)模型的高光谱遥感影像分类方法。该算法在半监督支持向量机的体系上加入未标记样本来辅助构建核矩阵,从而获得更优异的分类器,在小样本的基础上提高分类精度。试验结果表明,本文方法的分类精度好于传统方法,并且稳定性良好。
|
关 键 词: | 高光谱遥感影像 S3VM模型 未标记样本 半监督分类 |
Classification of Hyperspectral Remote Sensing Image Based on S3VM Model |
| |
Abstract: | The traditional hyperspectral remote sensing image classification is limited by the number of training samples ,so it is difficult to obtain the better classification results .This paper proposes a hyperspectral remote sensing image classification method based on semi-supervised support vector machine of clustering kernel .The method constructs a kernel matrix to obtain more excellent classifier by semi-supervised support vector machine and unlabeled sample ,and improves classification accuracy based on small sample .The experimental results show that the classification accuracy of this method proposed in this paper is better than the traditional method ,and has good stability. |
| |
Keywords: | hyperspectral remote sensing image S3VM model unlabeled sample semi-supervised classification |
本文献已被 CNKI 万方数据 等数据库收录! |
|