摘要
为有效地提高基于散射模型的非监督分类的分类精度,引入了Freeman三分量模型的改进模型-Yamaguchi四分量模型,并将该模型与威沙特距离模型结合起来。给出了基于四分量模型和威沙特距离的非监督分类、聚类算法及其实现流程。对AIRSAR数据集中的Flevoland图像选取了7个均匀程度不同的区域,进行了定性的、定量的实验,实验结果表明,新的分类、聚类算法能够显著的提高分类图的分辨率、更加清晰的表征地物的细节。该方法能够较大地提高均匀区域的分类精度。
To improve the accuracy of unsupervised classification based on scattering models, the four-component Yamaguchi model is introduced, which is a revised version of three-component Freeman model. Then, the four-component model is com- bined with the Wishart distance model. The new proposed algorithm of clustering is put forward whereafter and the procedure of this new method is listed. In experimentation, seven areas of various homogeneities are singled out from the Flevoland sample image in AIRSAR dataset. Qualitative and quantitative measure is performed. The experimental results show that the resolution and details are remarkably upgraded by the method. The accuracy of classification in homogeneous areas is increased a lot too.
出处
《计算机工程与设计》
CSCD
北大核心
2013年第7期2436-2440,共5页
Computer Engineering and Design
基金
国家自然科学基金项目(61202269)
广东省中国科学院全面战略合作基金项目(2011B090300041)
广东省教育部产学研合作基金项目(2011B090400430)
关键词
极化合成孔径雷达
非监督分类
分解模型
威沙特距离
四分量模型
polarimetric synthetic aperture radar
unsupervised classification
decomposition model
Wishart distance
four-com-ponent model
