基于极化SAR数据的图像分类识别算法  

Image Classification and Recognition Algorithm Based on Polarized SAR Data

作　　者：张军[1] 王勇[1] 郝万兵 杨磊[1] 郭鹏程[1] 

机构地区：[1]西安电子工程研究所,西安710100

出　　处：《火控雷达技术》2015年第4期55-59,64,共6页Fire Control Radar Technology

摘　　要：利用公开的全极化SAR数据,研究基于SAR图像的检测、极化分解和识别算法。首先根据四个线极化通道合成伪彩色图像,从而对场景进行初步认知。利用一维距离像分析全极化各通道的信噪比强度,通过对目标进行Pauli分解得到目标的奇次散射分量和偶次散射分量,从而完成对海杂波、建筑物和舰船的相干分量的研究。Detection,polarization decomposition and recognition algorithms based on SAR image is studied by using published full polarization SAR data. Firstly,a pseudo color image is synthesized according to four linear polarization channels so as to cognize a scene preliminarily. The signal-to-noise strength of each full polarized channel is analyzed by using one dimensional range profile; Pauli decomposition for the target is performed to obtain odd and even scattering components of the target so as to accomplish study on coherent component of sea clutter,buildings and ships.

关 键 词：全极化SAR数据 Pauli分解 伪彩色图像 

分 类 号：TN957.52[电子电信—信号与信息处理]