CAJ | 학술논문

给出了一种结合相位相关配准算法和迭代反向投影的彩色图像超分辨率重建算法。相位相关算法是一种利用频域信息估计图像之间偏移量的快速算法。在HSV颜色空间下利用低分辨率图像间的亚像素偏移量进行迭代反投影，保留图像的 H通道只对S通道和V通道进行迭代反投影。实验证明重建后的彩色图像与原始图像色调保持一致，并且很好地重建出图像的细节和纹理，避免了经过迭代之后造成的图像过度的平滑和颜色失真。
급출료일충결합상위상관배준산법화질대반향투영적채색도상초분변솔중건산법。상위상관산법시일충이용빈역신식고계도상지간편이량적쾌속산법。재HSV안색공간하이용저분변솔도상간적아상소편이량진행질대반투영，보류도상적 H통도지대S통도화V통도진행질대반투영。실험증명중건후적채색도상여원시도상색조보지일치，병차흔호지중건출도상적세절화문리，피면료경과질대지후조성적도상과도적평활화안색실진。
In this paper ,a super resolution technique which combines the phase correlation and iterative back projection (IBP) algorithm is proposed .Phase correlation is a method of image registration ,and a fast frequency‐domain approach is used to estimate the relative offsets between similar images .IBP algorithm is processed using the HSV color space .Hue is remained the same ,only saturation and value are used in iterative process .Experimental results demonstrate that this algo‐rithm can reconstruct the high resolution image with texture structures and details .Furthermore ,the high resolution image remains the same color tone as the original low resolution images and avoids the cross‐color and over‐smoothing made by the iterative process .