本文已被：浏览 659次 下载 1968次
投稿时间：2006-04-26 修订日期：2007-01-04
投稿时间：2006-04-26 修订日期：2007-01-04
中文摘要: 利用卫星图像对各种云型进行识别在大气科学领域具有重要意义，为了深入了解云分
类过程中逐个修改聚类和模糊聚类对各种云型的识别能力，采用极轨卫星EOS/MODIS图像资
料和静止卫星GMS 5图像资料，在样本采集和特征提取的基础上，选择不同的光谱或纹理特
征对两种分类器的分类性能进行测试和对比分析。结果发现，不管采用哪种图像资料，提取
哪些特征量，逐个修改聚类的平均分类准确率总体上略高于模糊聚类。但就两种分类器对各
种云型的识别能力而言，模糊聚类对低云和高云（如层云、薄卷云、密卷云、卷层云、积云
）的分类准确率明显好于逐个修改聚类，而逐个修改聚类对积雨云的分类准确率稍高于模糊
聚类。从各类别间混判的情形来看，积雨云和高中低混合云、低云之间及卷云子类之间混判
的情形较多，模糊聚类与逐个修改聚类相比，混判的类别增多，相对比例减少。
Abstract:In order to profoundly understand abilities of two classifiers—step
wise cluster and fuzzy cluster in the cloud classification techniques, both EOS/
MODIS and GMS 5 data set are used, spectral or textural features are drawn from
samples randomly to identify various cloud/surface. The results show that the st
epwise cluster gives higher accuracies than fuzzy classifier on the whole. With rega
rds to discriminating diverse cloud/surfaces, fuzzy cluster demonstrates its hig
her accuracies than stepwise cluster on the classes having similar characteristi
cs such as stratus, cumulostratus and cumulus; while stepwise cluster has better
capabilities of distinguishing cumulonimbus and surfaces. As far as misclassifi
cation of cloud/surfaces, fuzzy cluster tends to show lower accuracies in more m
isclassified classes.
文章编号： 中图分类号： 文献标志码：
基金项目:
引用文本：
朱亚平,刘健文,白洁,2007.云分类中逐个修改聚类和模糊聚类分类性能的对[J].气象,33(2):15-21.
Zhu Yaping,Liu Jianwen,Bai Jie,2007.A Comparative Study on Stepwise Cluster and Fuzzy Cluster in Cloud Classification Techniques[J].Meteor Mon,33(2):15-21.
朱亚平,刘健文,白洁,2007.云分类中逐个修改聚类和模糊聚类分类性能的对[J].气象,33(2):15-21.
Zhu Yaping,Liu Jianwen,Bai Jie,2007.A Comparative Study on Stepwise Cluster and Fuzzy Cluster in Cloud Classification Techniques[J].Meteor Mon,33(2):15-21.