基于物理散射模型的全极化SAR图像增强滤波算法  

Improved scattering-model-based speckle filter of polarimetric SAR data

作　　者：单子力[1] 谷宏志[1] 陈金勇[1] 王超[2] 

机构地区：[1]中国电子科技集团公司第五十四研究所,石家庄050081 [2]中国科学院遥感与数字地球研究所,北京100094

出　　处：《计算机应用研究》2013年第10期3152-3154,共3页Application Research of Computers

基　　金：国家"863"计划基金资助项目(2011AA120404)

摘　　要：针对传统的极化SAR滤波方法图像中城镇区域和植被区域地物在滤波中易被混淆,导致滤波后图像中地物边缘保持效果下降的问题,提出了一种增强的保持极化散射特性的滤波算法。利用一种增强的四分量极化分解方法获取更加精确的地物散射机制,并将散射机制信息引入滤波方法中,使滤波算法中像素的散射机制更精确。增强的四分量极化分解方法引入了极化SAR数据的定向角补偿技术、一种新的体散射模型以及两种散射功率限制条件,来改进Freeman-Durden分解的结果。理论分析和实验结果表明,改进后的方法获取了比传统的极化SAR图像滤波算法更加理想的计算结果。The conventional filtering algorithm tends to mix the urban areas with the vegetation areas for polarimetric SAR images. Therefore, the edges of the ground objects in the filtered images are usually not well reserved. To overcome this prob- lem, this paper proposed an enhanced filtering algorithm that could reserve the polarimetric scattering mechanisms, and applied an improved four-component polarimetric decomposition. The proposed filter included the orientation angle compensation of the scattering matrix, a new volume scattering model, and two power constraints to deal with the previous problems. Experimental results indicate that the proposed algorithm have acquired a better filtering effect than the conventional one for polarimetric SAR images.

关 键 词：全极化SAR图像 极化SAR图像滤波 增强的四分量极化分解 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TP301.6[自动化与计算机技术—计算机科学与技术]