期刊论文

A classification method of PolSAR data by combining polarimetric decomposition and multi-scale markov random field

中文

华中师范大学学报(自然科学版)

1000-1190

2014

48

3

436-441

10.19603/j.cnki.1000-1190.2014.03.026

国家自然科学基金项目（41301519）.

本校为其他机构

提出一种结合Pauli极化分解和多尺度马尔可夫随机场的极化SAR数据分类方法.该方法先将极化SAR数据进行Pauli极化分解来构造极化SAR特征数据,然后进行3层小波分解,最后进行尺度内和尺度间的迭代分类,获取极化SAR数据的分类结果.NASA/JPL数据的实验结果证明,无论是在地物轮廓保持性上还是地物细节描述性上,该方法效果都比较好.

In this paper,we proposed a classification method of polarimetric SAR data by combining the Pauli polarization decomposition and multi-scale markov random field.We first created the features of polarimetric SAR data by doing Pauli polarization decomposition.Then,these features were decomposed by three-layer Haar wavelet decomposition.Finally,iterative classifications were carried out in a layer and between layers to produce classification result of the polarimetric data.The experimental results by using NASA / JPL data show that the proposed method have better performance both on keeping the c...