基于Cloude目标分解的极化InSAR地物分类方法  

Polarimetric InSAR Object Classification Method Based on Cloude Target Decomposition

作　　者：钱方明[1,2] 楼良盛[1,2] 刘薇[1,2] 张笑微[1,2] 张昊[1,2] 

机构地区：[1]西安测绘研究所,陕西西安710054 [2]地理信息工程国家重点实验室,陕西西安710054

出　　处：《测绘科学与工程》2016年第1期31-35,共5页Geomatics Science and Engineering

摘　　要：本文针对利用Pol—InSAR数据进行地物分类的问题，提出了一种基于cloude目标分解的Pol—InSAR地物分类方法。该方法首先对Pol—InSAR数据进行预处理，抑制噪声影响，然后采用Cloude目标分解结果和极化相干性信息确定初始类别，最后对初始类别利用Wishart分类器进行聚类分析得到分类结果。实验结果表明，该方法在地物类型单一区域分类精度较高，在地物类型多样区域分类精度相对较低。An object classification method based on Cloude target decomposition for polarimetric InSAR （ Pol - InSAR ） is proposed in this paper to resolve the classification problem using Pol - InSAR data. The classification principles and procedure are described and designed. Firstly, the Pol - InSAR data is pre - processed to reduce noise impact, and then the Cloude target decomposition result and polarimetric coherence information are used to confirm the initial classification result. At last, cluster analysis is conducted with Wishart classifier to obtain the final classification result. The experiment shows that the classification precision is higher in areas of single - type object and relative lower in areas of various objects.

关 键 词：极化干涉合成孔径雷达 地物分类 Cloude目标分解 

分 类 号：P231[天文地球—摄影测量与遥感]