摘要
混合分类既可以改进纯粹的监督分类或非监督分类过程的效率,又可以提高纯粹的监督分类或非监督分类过程的精度。介绍一种新的遥感混合分类方法———控制聚类法(guided clustering)。首先对该方法的原理进行阐述,然后利用该方法对重庆市具有代表性的山区酉阳县的遥感影像进行分类实验。实验结果表明,这种方法能有效的提高分类精度,特别适合我国南方山区土地覆盖信息的提取。
The hybrid classification can improve the efficiency of process on supervised classification or unsupervised classification. First the theory of guided clustering was described. Then this method was developed to make a classification experiment using TM imagine in the representative mountain area-Youyang. The results showed that the guided clustering can increase the classification accuracy,especially suitable to apply in this south mountain area.
出处
《水土保持研究》
CSCD
北大核心
2006年第3期235-237,共3页
Research of Soil and Water Conservation
关键词
遥感
控制聚类法
山区
混合分类
remote sensing
guided clustering
mountain areas
