中文摘要：

本文结合Freeman分解和子孔径分析,提出一种新的极化SAR图像分类算法。该方法首先利用子孔径分解,产生不同方位观察角度下的子孔径图像,再利用Freeman分解对各个子孔径图像提取三种散射机理成分的功率,平均后对类别进行细分,最后使用Wishart统计分类器对类别进行分类划分得到最终结果。该方法考虑了极化散射机理在不同方位观察角度下的变化,能够取得较好的分类效果,能够保存主要极化散射特性的纯度,同时还可以动态地设定分类类别数。最后利用EMISAR获取的极化SAR数据进行了仿真,验证了该方法的有效性。

英文摘要：

In this paper,a new method,which combines the Freeman decomposition with sub-aperture analysis,was used to classify the PolSAR images.Firstly,sub-aperture decomposition was used to produce datasets,which corresponded to different observed azimuths.Secondly,Freeman decomposition was employed to abstract the powers of the mechanisms.After powers averaging,more clusters were produced.What is more,Wishart classifier was used to obtain the final result.The change of polarimetric scattering in different observed azimuths was considered in this method,so better classification result could be obtained.It also could preserve the purity of dominant polarimetric scattering properties and obtain a dynamic number of clusters.Finally,the EMISAR data was applied in classification experiment,which was used to validate the feasibility and efficiency of the algorithm.