一种自适应选择酉变换的PolSAR三分量分解方法  

Polarimetric SAR Three Components Decomposition with Adaptive Selection of Unitary Transformations

作　　者：丁仕军 DING Shijun(Land Surveying and Mapping Institute of Shandong Province,Jinan 250102,China)

机构地区：[1]山东省国土测绘院,济南250102

出　　处：《遥感信息》2021年第5期69-73,共5页Remote Sensing Information

摘　　要：基于模型的目标分解方法是全极化合成孔径雷达图像解译的重要方法,可以得到目标的散射机制信息,是进行目标分类、目标识别的基础。三分量分解方法是一种基于模型的目标分解方法,具有实施简单、物理意义明确的优点。基于崔一等提出完全利用极化信息的模型分解方法,文章对剩余矩阵进行自适应选择的酉变换,达到剩余矩阵更加匹配面散射模型和二面角散射模型的目的。首先,对自相关矩阵进行非负特征分解得到体散射分量,得到剩余矩阵;然后,对剩余矩阵进行自适应酉变换,以使交叉极化散射分量最小,更加匹配面散射和二面角散射模型;最后,通过实验验证所提出的三分量模型分解方法在极化散射机制解译方面的有效性和优越性。As an important interpretation method for polarimetric synthetic aperture radar(PolSAR)data,model-based target decomposition method derives the targets’scattering mechanism,and is the basis of target classification and target recognition.As one of the model-based decompositions,three-component decomposition has the advantages of easy implementation and clear physical meanings.Based on the complete model-based decomposition by CUI,adaptive selection of unitary transformations are implemented on the remaining coherency matrix,and the remaining coherency matrix would match better with the models of surface scattering and double-bounce scattering.First,with one nonnegative eigenvalue decomposition implemented,volume scattering component is derived,and the remaining coherency matrix is obtained.Then,adaptive unitary transformations are implemented on the remaining matrix,and as a result cross-polarization scattering component is minimized,which matches better with the surface scattering and double-bounce scattering models.Last,experiments validate the effectiveness and superiority of the proposed method.

关 键 词：全极化合成孔径雷达 三分量分解 酉变换 模型分解 非负特征分解 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]