论文部分内容阅读
将贝叶斯反演理论应用于综合孔径辐射计图像反演,其基本思想是将亮温分布和可见度视为随机变量,将图像反演转换为贝叶斯推断问题.基于这一思想,分析了亮温分布的先验信息和可见度的统计特性.鉴于热辐射信号具有高斯特性,建立了亮温分布的高斯先验模型,并采用期望最大算法估计该先验模型的未知参数.可见度采用多元高斯随机变量表示,并考虑了可见度的互相关性.贝叶斯反演法能够充分利用可见度和先验信息的统计特性.仿真和实验结果表明:贝叶斯反演法能有效提高综合孔径辐射计的成像性能.
The Bayesian inversion theory is applied to the inversion of synthetic aperture radiometer image, its basic idea is to consider the brightness temperature distribution and visibility as a random variable, the image inversion into Bayesian inference problem.Based on this idea, analysis The statistical properties of the prior information and the visibility of the brightness temperature distribution are obtained. In view of the Gaussian characteristics of the thermal radiation signal, a Gaussian prior model of the brightness temperature distribution is established and the unknown parameter of the prior model is estimated by the maximum expectation algorithm. Gaussian random variables and consider the cross-correlation of visibility Bayesian inversion method can take full advantage of the statistical properties of visibility and prior information.The simulation and experimental results show that the Bayesian inversion method can effectively improve the comprehensive aperture radiation The imaging performance of the meter.