基于稀疏先验的光学及SAR图像的分辨率增强统一框架

Unified frame based on sparse prior for optical and SAR image resolution enhancement

研究了基于稀疏型先验的光学及SAR图像的分辨率增强的统一框架。在定义稀疏型先验的基础上, 分析了稀疏型先验与分辨率增强的关系,并从参数估计的角度解释了为什么稀疏型先验的合理利用可以实现分辨率增强。其次给出了光学及SAR图像的统一观测模型。再次给出了光学及SAR图像先验模型的统一描述。最后给出了光学及SAR图像分辨率增强的目标函数构造的统一形式。并结合光学图像和SAR图像,在不同数据域的先验信息,分别构造各自的目标函数,实现图像分辨率增强。

The unified frame based on sparse prior for optical and SAR （synthetic aperture radar） image resolution enhancement is studied. First, we define the sparse prior, analyze the relationship between the sparse prior and resolution enhancement, and explain by parameter estimation theory. Second we unify the observation model for both optical and SAR image. Thirdly we unify the prior description. Finally the unified object function is constructed for both optical and SAR images. The examples for both optical and SAR images are presented, in which the object function are based on the sparse prior in different data domain respectively.