基于极化特征识别的地貌识别策略  

Landscape Recognition Strategy Based on Polarization Characteristics

作　　者：谢景海 孙密 郭嘉 贾祎轲 卢诗华 苏东禹 XIE Jing-hai;SUN Mi;GUO Jia;JIA Yi-ke;LU Shi-hua;SU Dong-yu(Economic and Technical Research Institute of Jibei Electric Power Co.,Ltd,Beijing 100038,China)

机构地区：[1]国网冀北电力有限公司经济技术研究院,北京100038

出　　处：《火力与指挥控制》2020年第6期73-79,共7页Fire Control & Command Control

基　　金：国家重点研发计划基金资助项目(2017YFC0804111)。

摘　　要：为满足其近感测量精度,需对探测范围内的地貌环境进行有效识别。提出一种基于最优投影的地貌识别算法,为现代空域近感平台提供必要的地貌识别能力。定义了利用目标散射矩阵特征值构成的特征参数,且该参数与探测信号强度、角度等参数均无关,仅相关于探测区域内的地形结构特征。并结合Krogager关于目标散射矩阵分解的相关思想,建立带参数的标准散射体旋转不变矩阵,将探测区域内的地貌极化散射矩阵投影至标准体,得到目标散射矩阵在每一标准体的权值,并通过最优化投影进一步判定该地貌属性,从而达到识别的目的。In order to meet measurement requirements,the landscape environment in the detection range must be recognized,therefore,a novel landscape recognition algorithm based on optimal projection has been proposed,in this paper,to provide new generation projectile with necessary recognition ability.The characteristic values of target scattering matrix eigenvalues are defined,and it has no directly relation to power and angle of detection signal.It is only related to topographic features of landscape in detection range.Meanwhile,a series of standard scatter rotation invariant matrixes with uncertain parameters have been constructed according to the ideas,decomposition of target scattering matrixes,from Krogager.The geomorphological polarization scattering matrix of detection range is projected into standard scatter projects and relevant weight values are figured to judge the landscape attributes,so,our goal,recognize landscape,can be realized.

关 键 词：极化信号 最优投影 空域平台 地貌识别 旋转不变矩阵 

分 类 号：TN953[电子电信—信号与信息处理]