基于不可控微扫描的高分辨力图像重构方法

针对存在不可控微位移的序列低分辨力图像,提出了一种基于2×2不可控微扫描的高分辨力图像重构方法,采用投影法估算出低分辨力图像LR的微位移量,采用基于泰勒级数展开的重构算法由4帧图像重构高分辨力图像.模拟表明：该算法能精确地估计序列图像的帧间位移量并取得较好的重构效果,且算法简单有效可行,处理量小,有利于实现快速处理,在一定程度可满足同步轨道卫星、公安监视等领域对高分辨力成像处理的需求.  

High Resolution Image Reconstruction Basedon Uncontrolled Microscanning

Aiming at a series of Low Resolution( LR ) images with uncontrolled micro displacements, an algorithm based on 2×2 uncontrolled microscanning is represented to reconstruct the High Resolution(HR) image. It uses the algorithm of Linear Interpolation Projection based on Shift Estimation(LIPSE) to estimate the micro displacements of the  LR  images, and then reconstructs the  HR  image from the four  LR  images by using the reconstruction algorithm based on Taylor series expansion. Simulation and practical image processing demonstrate that this algorithm can estimate displacements accurately of serial images and get good reconstruction effect, and the algorithm is simple but effective with small processing load, so it is very possible to achieve fast processing. The method is hopeful to meet the demand of  HR  imaging in the fields of synchronous satellite and public surveillance.