极化SAR图像去相关目标检测改进算法

发布时间：2018-05-23 15:51

本文选题：极化合成孔径雷达图像 + 极化匹配滤波　；参考：《计算机工程》2016年09期

【摘要】：针对极化合成孔径雷达(SAR)图像为多通道数据的特点,提出一种改进的极化SAR图像去相关目标检测算法。分析经典极化匹配滤波检测量在求取最优权值及后续统计分布推导中存在的问题,利用实测数据验证不同地物极化SAR数据通道之间的相关性。根据二维高斯分布的去相关理论,对极化匹配滤波检测量数据通道之间进行去相关,得到新的去相关极化匹配滤波检测量,该检测量在最优权值的求取上更具一般性。经过去相关处理后的极化匹配滤波检测量的各通道数据不存在相关性,且满足独立分布的复高斯随机变量,使后续的统计分布推导更为严格。实验结果表明,该算法能够有效区分目标与杂波,具有较高的检测率,且虚警目标点少。
[Abstract]:An improved decorrelating target detection algorithm for polarimetric synthetic aperture radar (SAR) images is presented in this paper, which is based on the multi-channel data of polarimetric synthetic Aperture Radar (SAR) images. This paper analyzes the problems of classical polarization matched filter measurement in obtaining the optimal weight value and the derivation of subsequent statistical distribution, and verifies the correlation between different ground object polarization SAR data channels by using the measured data. According to the theory of de-correlation of two-dimensional Gao Si distribution, the data channels of polarization matched filter detection are de-correlated, and a new de-correlation polarization matched filter detection is obtained, which is more general in the calculation of optimal weights. The data measured by polarization matched filter after de-correlation processing have no correlation, and satisfy the independent distribution of complex Gao Si random variables, which makes the subsequent statistical distribution derivation more strict. The experimental results show that the algorithm can effectively distinguish the target from the clutter, has a high detection rate, and has fewer false alarm points.
【作者单位】：河南工程学院计算机学院;
【分类号】：TN957.52

【相似文献】