摘要
本文首先通过分析全极化干涉SAR条件下的散射机理,分别计算出来自地表层和植被层的散射功率,并进一步通过确定像素在功率平面上的位置来区分目标,实现对地物的分类。在此基础上,为了充分利用H/α非监督分类结果的有用信息,对基于H/α分类和基于散射功率分类得到的分类集进行组合,并通过复Wishart迭代分类方法进行聚类,得到最终的SAR图像分类结果。本文详细阐述了该方法的原理和实施步骤,并对SIR-C/X-SAR的L波段实际全极化干涉数据进行分类实验,可知该算法无论在分类精确度上还是在迭代速度上,都有较高的性能。
In this paper, the scattering mechanism in the condition of polarimetric interferometrie SAR was firstly analyzed, and then the scattering power from ground layer and vegetation layer were calculated respectively, finally the classification was realized by judging the position of every pixel in power plane. For utilizing the useful information of Lee＇s classification, two classified sets of Lee＇s and our algorithm were combined, and then the classified sets were aggregated to get the last results by the complex Wishart iteration classification. In the paper, the algorithm was described in detail and the eontrastive experiment was done with SIR-C/X-SAR L band polarimetrie interferometric SAR images. The experiment results indicated that the classification accuracy of the method in the paper was higher and the iterative speed was faster.
出处
《测绘科学》
CSCD
北大核心
2010年第4期34-37,共4页
Science of Surveying and Mapping
关键词
极化干涉SAR
功率
散射
迭代
非监督分类
polarimetic interferometric SAR
power
scatter
iteration
unsupervised classification
